S2E17 - Red Means Stop

Have you ever sat at a red light at 2 AM with no traffic in any direction and waited anyway? Have you ever rolled through that same red light 2 AM and felt vaguely guilty about it?
Of course you have. The traffic light is the most obeyed command in human history. Rarely enforced (unless you're in the UK like Marc). No officer in sight. Just a coloured light on a pole, and a near-universal agreement to stop when it's red and go when it's green. 
This episode traces the humble traffic signal from the gas-lit lantern that exploded outside the Houses of Parliament in 1868 (yes, exploded, three weeks in) to the adaptive AI systems that watch real-time traffic and adjust timing in milliseconds. Along the way: railroad colour conventions, William Potts in Detroit and Garrett Morgan in Cleveland, the political question of whose green is longer, the inductive loop that can't see your bicycle, and the moment where you discover that the colour you grew up calling yellow is officially called amber once you cross an ocean.
Ride along with Marc and Renee through another look at a technology that became infrastructure as it spread beyond its humble beginnings.

Join Renee and Marc as they discuss tech topics with a view on their nostalgic pasts in tech that help them understand today's challenges and tomorrow's potential.

email us at nostalgicnerdspodcast@gmail.com

Come visit us at https://www.nostalgicnerdspodcast.com/episodes or wherever you get your podcasts.

Transcript

SPEAKER_01 0:00

I want to talk about the most obeyed command in human history. Not a law, exactly. Not an instruction from a person in authority, just a colored light on a pole at an intersection. And the near universal human agreement to stop when it's red and go when it's green. Think about that. Think about what that actually is. There's no enforcement mechanism at most traffic lights. Well, now they put cameras there. I mean, we have a mirror. They put cameras there now. And then you get sued and they take them back. I don't know. There's no camera watching you at the majority of intersections. There's no officer president. It's a light and it's a shared social understanding so deeply embedded that most people stop without consciously deciding to. The way you pull your hand back from a hot stove before the pain signal registers, the traffic light is one of the most successful pieces of behavioral technology ever deployed. And we installed it on practically every urban corner in the developed world before anyone had a clear theory of why it worked. I have not. I have not. And we have roundabouts here, so that's hairy all by itself. I'd rather have a traffic light, to be honest with you.

SPEAKER_02 1:19

I think there's two traffic lights in the whole city of Cairo. Lines on the road, there are suggestions. There's suggestions of the city.

SPEAKER_03 1:29

Yeah, places like that.

SPEAKER_02 1:31

Yeah, yeah. The traffic light is also, from a systems engineering perspective, one of the most consequential pieces of urban infrastructure in existence. Yeah, that's totally true. The timing of signals across the city's intersection net intersections network determines how efficiently the road system moves traffic, which determines commute times, fuel consumption. You know, right turn on red in the US, that's a fuel consumption thing, Jimmy Carter. Emergency response times and air quality. A city with uh see, right turn on red, you know, air quality too. So you're not sitting there idling the whole time. Yeah, a city with well-optimized signal timing moves meaningfully more vehicles than one without it. Uh, on the same physical road infrastructure with no additional construction. Traffic signal optimization is essentially free capacity expansion. And most cities are leaving significant amounts of it on the table.

SPEAKER_01 2:29

So we're gonna go from the first traffic signal, which was a gas-lit rotating lantern operated by a police officer in London in 1868, to adaptive AI systems that watch real-time traffic flow and adjust the signal timing in milliseconds without any human involvement. The arc covers about 150 years and involves railroad engineering, the first electric traffic light in Cleveland of all places, and a black inventor in Detroit who held the patent on the modern three-color signal, the the specific physics of why green means go and red means stop, and the moment in the 1970s when traffic engineers first started treating the city's signal network as a single computational system rather than a collection of independent intersections. It's a history that's hiding in plain sight at every corner you have ever waited on.

SPEAKER_04 3:23

Three seconds is all like you look at me like you know what I'm trying to say. Break a little bit, get a little bit. Either way, you answer something I'm never quite said on.

SPEAKER_02 3:52

So the traffic light also has a governance dimension that doesn't get discussed enough. Governance, Renee's favorite topic. Uh every every timing decision embedded in a traffic signal is implicitly a political decision about whose time and whose movement the city prioritizes. A signal that gives a long green phase to a major arterial road, moving commuters from suburbs to downtown and a short green to the cross streets serving a residential neighborhood has made a choice about whose travel matters more. The signal timing is transportation policy expressed in seconds, and it operates at every intersection in the city simultaneously, mostly without anyone noticing the choices being made.

SPEAKER_01 4:33

No, no, we notice. We notice.

SPEAKER_02 4:38

I always wonder like, like, who's whose job is that one? Like, you know, and if and if you knew that Bill, uh the the traffic light engineer guy down at the city, you know, was the guy and he's your neighbor, and he's like, Yeah, would he turn on the one at the bottom of the street?

SPEAKER_01 4:55

Yeah, yeah.

SPEAKER_02 4:56

Bill, why the heck is that light so short? I you know, it takes me forever to get through that light. Poor Bill. To understand where the traffic light comes from, you you have to understand the railroad signal system that preceded it. The traffic light is an application of existing railroad technology to a new problem. By the mid-19th century, railroads have developed a sophisticated system of colored light signals for managing train movements on shared tracks. The I you know what, we actually have one of those red lanterns. And it's really cool. Yeah, the logic is simple and robust. Med red means stop because red is the color of danger. A convention drawn from maritime signal flags and reinforced by its visibility at distance. Green means proceed, or in some early systems, caution. White or clear means the line is safe. The color coding is standardized across railroad operators so that any engineer can interpret any signal without knowing the specific practices of that particular line.

SPEAKER_01 5:59

And the reason red means danger in the first place is genuinely interesting. And it goes back further than railroads. Red's been associated with warning and prohibition across an enormous range of cultures and historical periods, and there's competing theories about why. The most compelling one is physiological. Red light is the longest wavelength in the visible spectrum, which means it scatters least in fog, rain, and atmospheric haze. The red signal is visible at greater distances and in worse weather than any other color, which is why I wear red lipstick. So you can keep see me coming in any weather. The association of red with danger may have been re may have been reinforced across generations precisely because red warnings were the ones people could actually see in time to respond to. The color coding that feels arbitrary and conventional is possibly rooted in practical physics.

SPEAKER_02 6:54

Possibly. Possibly. Yeah, I think we'll get into it maybe in the other a little a little later, but like some of the colors, it's different, you know, in different parts of the world. It's in some places opposite, which is kind of interesting. So John Peake Knight, a railway signal engineer, proposes the first traffic signal for road use in 1868 to address a specific problem at the intersection of Bridge Street and Great George Street in Westminster, London, near the Houses of Parliament. I know that intersection. The intersection is chaotic, horse-drawn carriages, pedestrians, and other traffic coverage without any coordination mechanism. Knight's mechanist proposal is to install a version of the railway semaphore signal he he had already used professionally, a mechanical arm that indicates stop or caution, supplemented. I know they go swinging. You know, this is like Ding Bing. Yeah, exactly. So I was thinking about this, and unless you know people who are listening have seen one, you know, in action, the the reference they'll probably have is from a Looney Tunes cartoon. Because in Looney Tunes or, you know, like early, you know, or mid-century cartoons, they still used the swinging semaphore style signals in a lot of the visual gags. So that would be that would be how you would maybe you know know that. So but basically that mechanical arm indicates stop or caution, supplemented at night by a gaslit lantern with red and green lenses that rotates to show the appropriate color. The signal is operated manually by a police officer stationed at the intersection. Boy, that's a rough job.

SPEAKER_01 8:35

I know, right? Like the installation goes live on December 9th, 1868, and it works for about three weeks before it explodes. The gas company, the the gas supply to the lantern leaks and the lantern ignites, and the police officer operating it is burned badly enough to require hospitalization in 1868. Oh, how horrible would that be? The project is abandoned, and London goes back to uncontrolled intersections for the next several decades. This is, on reflection, not an encouraging start for a technology that now governs movement at hundreds of thousands of intersections worldwide, but it does establish the color coding and the basic operating concept that all subsequent traffic signals inherit. Like right? Like I almost thank God for that poor cop. Like an unsung hero in history.

SPEAKER_02 9:28

Red really meant caution. Uh I almost did the song about that. Like, but I thought, oh, that would be too one, too obscure. And it'd be like too kind of like, you know, right on the nose. So, anyways, the first electric traffic signal appears in Cleveland, Ohio, in 1914, installed by the American Traffic Signal Company at the corner of East 105th Street and Euclid Avenue. Does every every town in America have Euclid Euclid, yes. Yeah. Yeah. Uh it uses red and green lights and an audible warning buzzer operated by a police officer in a booth at the intersection. I you know, I think I wonder if some of these booths still exist. The electric quick to the Google. I know, right? The electric version solves the explosion problem of the gas uh lantern and allows for for more reliable and consistent operation, but it still requires a human operator. The fundamental limitation of the manually operated signal is that the operator has to make judgment calls about timing based on what they can observe from their position. And they can't optimize for conditions they can't see.

SPEAKER_01 10:40

So Garrett Morgan, a black inventor and entrepreneur in Cleveland, patents a mechanical three-position traffic signal in 1923. And his contribution is often unrecognized in histories that focus on later developments. Morgan's design is a T-shaped pole with a movable arm that pick can be set in three positions: stop, go, and an all-stop interval that brings traffic in every direction to a halt before the signal changes. The earlier two-position red and green system created a dangerous moment of ambiguity at the end of the transition. So drivers approaching on the green who would see it about to change, like they'd go faster. While cross traffic was anticipating their green, so they'd go. And then that the same principle applied to the electric three color signals later in the decade. The becomes a yellow phase, like you just go faster, right? It's yellow, go faster. And then if you get there, right as it's starting to turn red, you convince yourself, no, dude, that was orange. I'm good. I'm good. Like no cop could pull me over. That was orange. It was definitely still sort of yellow. Morgan sells his patent to GE, General Electric, reportedly for 40, for$40,000. I mean, a lot at the time. 1923 money. Okay, but if he invested that in the market, God help him, because it was gone the next week.

SPEAKER_02 11:55

Right?

SPEAKER_01 11:56

Yeah.

SPEAKER_02 11:57

Well, he had six years of, you know, good investments, hopefully. William Potts, a Detroit police officer, independently develops and installs the first four-way, three-color traffic signal in Detroit in 1920. Three years before Morgan's patent, the the historical record on the relative contributions of Morgan and Potts is, you know, contested and complicated by the fact that Morgan, as a black inventor in the early 20th century, operated in an environment where patents held by black inventors were frequently challenged, circumvented, or simply ignored by manufacturers who preferred to avoid the royalty obligation. So of course they did. Yeah, of course they did.

SPEAKER_01 12:35

Of course they did.

SPEAKER_02 12:41

It here in the UK, we call that amber, just so you know. Oh, amber. Amber, yeah. With an amber yellow warning phase. See, but you know, because I grew up in the States, yellow. Yeah, yellow. It's a yellow light. Yeah, exactly. But here everybody says amber. Anyways. Yeah. There's there's a as I said, there's contention. Morgan and Potts both contribute to that sort of convergence there. You know, as I said, that convention varies by by country. I called it yellow. UK regulations and signals call call it amber, and there's a four-phase process. So it goes green, amber, red. So that's normal. But then there's a red and amber together before returning to green. And the red amber phase warns drivers that change is imminent and gives them time to release the brake. I think it's because they do the start-stop thing now. You know, the auto-stop. So when it's yellow, when it turns, you know, it's red, solid red, and then it goes red amber, red yellow, whatever. Then you can let off the brake and let your car start to do the automatic start. Oh, I see. Anyways, but this existed longer, longer than that. Anyways. But it gives you time to release the brake. Continental Europe signals follow the same pattern as the UK. And an American three-phase sequence is just, you know, one regional variant among several around the world.

SPEAKER_01 14:01

Next act. The early traffic signal is a standalone device. It cycles through its phases on a fixed timer, the same sequence of green, yellow, red in the same durations, regardless of whether there's 10 cars in the intersection or one or none. It's just, it's going to go through that process. This is better than nothing and considerably worse than optimal, right? A fixed time signal gives 60 seconds of green to a major arterial road and 30 seconds to a side street, is making a judgment about the typical traffic ratio of that intersection, calibrated at the time of installation, that will execute unchanged until someone physically reprograms the controller. If the neighborhood changes, if a new development goes in nearby, if a traffic pattern shifts seasonally, the signal doesn't adapt. It just keeps sight. Westwood Boulevard in LA. You get down toward like Pico and you'll sit there forever in that turning lane. You'll sit there forever. Forever.

SPEAKER_02 15:05

I think you you get to a certain point where like, you know, the the sensors, which we'll talk about in in a minute, but the sensors, like it doesn't matter how much traffic there is, you know, every every direction has a lot. So it just does what it wants to do.

SPEAKER_01 15:20

Right. Right. It'll just start you're not getting left because those guys keep taking up all the I c I I can't get left.

SPEAKER_02 15:31

Maybe that's why the UK has roundabouts, is because you know the guy it does move traffic better. It does move traffic, but also, you know, the guy he got a his arm, you know, the explosion at the signal. Oh, you guys are just gun shy with this stuff? Yeah. I mean today. It just ticks tick but by the by that time, it's like, well, we'll just do roundabouts. Well, anyways, the the coordination problem is where traffic engineering becomes a genuine systems challenge. A signal, a single intersection can be timed in isolation without too much difficulty. But you know what? Updating it, like you gotta update every single intersection. What a pain. Yeah. You know, thousands of intersections. And the timing decisions at each one interact with the decisions at every adjacent one. A driver who gets a green at one intersection arrives at the next intersection at a predictable time. And if the next signal is limited to give or is time to give green at exactly the moment, the driver moves through without stopping. This is the concept of a green wave or signal progression. And achieving it across a corridor or network requires that all the signals in the system be coordinated with each other. Um in Ontario, I remember when I was a kid, the signals, there was a sign that said speed limit like 45 or something like that. Signals timed to 40 miles an hour. So they purposely, you know, would encourage a slower speed because that was the that was the green wave.

SPEAKER_01 16:55

I was always convinced, yeah, I was always convinced you had to go at least six miles over the speed and I might catch the green wave. So here I am, like busting through intersections just so I can get there faster, so I can catch it before it turns yellow. I always thought it was faster. I'm just gonna keep believing that. In my heart, I'm gonna believe that. The first coordinated signal system goes into operation in Salt Lake City in 1917 with six signals along a single corridor, time to allow a continuous flow of traffic at a specific speed. The concept works, and the efficiency gain is immediately apparent. Vehicles moving through a coordinated progression stop less frequency, they consume less fuel, and they travel faster than the vehicles navigating, you know, uncoordinated signals. The challenge is that coordination requires central control, which means communication infrastructure connecting the signals to a central timing source, which in 1917 is a non-trivial engineering problem.

SPEAKER_02 17:55

They didn't have Wi-Fi.

SPEAKER_00 17:57

They did not have Wi-Fi, there was no RIDF, there was no, there's none of that.

SPEAKER_02 18:01

No Bluetooth, no.

SPEAKER_00 18:03

No Bluetooth, you gotta figure this out.

SPEAKER_02 18:05

You couldn't run the Cat 6, you know, no fiber connection on it.

SPEAKER_03 18:10

Fiber.

SPEAKER_02 18:11

No. Yeah. The electromechanical master controller systems that emerged through the 1920s and 30s use physical connections, telephone lines, or dedicated cables to synchronize signals. I bet you that that copper pipe was that copper is thick too. Right.

SPEAKER_01 18:28

It was like coax. Yeah.

SPEAKER_02 18:33

The master controller runs a cycle and and and the field signals lock their phases to it. The system works well for simple corridor progressions, uh, but becomes increasingly difficult to manage as the network grows and as planners try to optimize for multiple competing objectives simultaneously. You can time a corridor for maximum progression in one direction at a time. The perpendicular direction needs a different timing. Every coordination decision privileges some movements over others.

SPEAKER_01 19:03

This is complicated. You can see why they don't do it in Cairo. I mean, who would want to sit down and figure all this out for real? I mean, I guess you could give it to ChatGPT now, but come on now. The political dimension of signal timing that Mark mentioned at the top of the episode is m most visible in the history of pedestrian signal timing. For most of traffic signal history, pedestrian costing time has been determined by engineering formulas that calculate how long it takes an average adult to cross a street at an average walking speed. Now, I don't know. I find myself running across really long stuff. Else I get so so who is the guy? Was he like 10 feet tall? Yeah. Nine feet tall? Like I find myself, all right. The problem is that the population using the intersection is not composed of average adults. Bingo. Elderly, right? Elderly per pedestrians walk more slowly. Children walk unpredictably. Dogs just lay there when they're lazy. People with mobility impairments may need significantly more time than the formula assumes. Intersections timed for the average pedestrian systematically fail the lowest and most vulnerable pedestrians who are often the people most dependent on walking as a mode of transportation. The signal timing formula is not neutral. It expresses an assumption about who those streets are for, and they are not for old people, kids, or dogs.

SPEAKER_02 20:26

Or bicyclists, you know, or people who are five foot three.

SPEAKER_01 20:30

I it's not for me either.

SPEAKER_02 20:32

I mean, like the whole idea of average, you know, there, it just doesn't. It means like there's an awful lot of people below below the average, you know?

SPEAKER_01 20:42

So I And let's face it, like like but we're talking about the early 1900s. So this is a man. You're generally speaking, y'all are taller than us, right? Yeah. You definitely walk faster than we do, you're not wearing high heels. Like there's a whole thing about that, right? I mean, Christmas dummies didn't have you know female anatomy until like 1990s. Exactly.

SPEAKER_03 21:02

Yeah.

SPEAKER_02 21:02

Yeah.

SPEAKER_01 21:03

Well, I just yeah.

SPEAKER_02 21:05

I don't know if you wear heels in London, you know. I like if you you've walked on those streets, they're you know, they had cobbles and they get uneven. Like, yeah, forget about that. I wear tennis shoes there no matter what. Oh yeah, yeah. Yeah. But you you really should be targeting like 80 percentile, you know, not average. You know, just yeah, whatever. Agree. I'm sure it's all great now, right? With all this magical stuff. Perfect. Perfect. The vehicle detector is the technology that will make this all work. And that enables the transition from fixed time signals to actuated signals. The inductive loop detector, which is the standard technology from the s 1960s through today, is a loop of wire embedded in the pavement at or near the stop line. And you'll see these, they look like they're like like caulk or putty that they've yeah, circles or hexagons, you know, or or whatever that they've cut out of the ground. It detects a vehicle, you know, the presence of a vehicle. A simple actuated signal uses the detector to extend the green phase as long as vehicles are present up to some maximum duration rather than cycling on a fixed timer. This allows the signal to respond to actual traffic demand rather than assumed demand, which improves efficiency significantly at low volume or variable volume intersections.

SPEAKER_01 22:25

Check it out. I used to call it riding the O's. So if you were going down Westwood and you wanted to get in that tandem lane and you wanted that arrow not to be like a steady green light, you wanted the arrow, you would have to wait for that whole thing to fill up for it to turn because the O's were way at the back. They weren't up by the light. So I would just sit there way at the back. Everybody would cut in front of me to fill in that space, but I would just sit there because I knew I wanted that stupid turn signal to turn on. And I would just sit there and it would like green arrow. I'm like, we're out of here. Like, and because I'm still close enough, I'd make it through that light. It's like it's called riding the O's, people. You should do it more often. The inductive loop has a systemic blind spot that became a significant issue as the alternative transportation grew. It doesn't reliably detect bicycles.

unknown 23:12

Yeah.

SPEAKER_01 23:12

And occasionally the mini. It didn't realize that the mini, because it was so short, was taking up the whole space. I could sit there forever. It wouldn't know. Yeah, it was terrible because bicycles have too little metal mass to trigger the induct the inductance change reliably. The cyclist arriving at the red light with the loop detector might sit through multiple cycles without the signal ever detecting their presence because the signal doesn't know that they're there. This is a piece of infrastructure that is literally invisible to a significant and growing category of road user. And it persists because of the fix. Replacing loops with video detection or other technologies requires opening up the pavement and at every affected intersection, which is expensive and disruptive. So here's my advice, bicycle people. Just hit the walk button and get back on your bike and just wait for it to change.

SPEAKER_03 24:04

You're welcome.

SPEAKER_02 24:05

You hit the walk button. I've seen that. I have seen people do that. And now, have you seen the little the little delivery bots that get stuck at the I haven't seen one get stuck anywhere yet, but yeah, I see what it gets stuck and it flashes and it like says help, help. And it like asks people to press the press the button, you know? Like in here in the UK, the at least in London, the the signals for walk, they're not like you the button doesn't do anything. All it does is like it turns the light on that says walk. But they always they always cycle for pedestrians no matter what. But in other parts, you know, like not in London. Like I saw one in Leeds the other day when I was up there for for Ella. And and the and it was it was just literally stuck. And it nobody could, you know, nobody was there to press the button.

SPEAKER_01 25:00

So that's sad. It just sat there saying help. Yeah. Help me. Yeah.

SPEAKER_02 25:06

Anyways, we'll have to do an episode on delivery bots because I've I've spent a lot of time on those things. So that'd be fun. Yeah. Anyways, the first computerized traffic control systems arrived in the early 1960s when Toronto in the magical land of Toronto. Toronto. They install a computer controlled signal network covering a portion of the downtown street grid in 1963. The computer is able to implement signal timing plans that would be impractical to execute with electromechanical controllers and to switch between timing plans based on the time of day or detected conditions. The Toronto system demonstrates that centralized computational control of a signal network is technically feasible and produces measurable efficiency improvements over manual coordination.

SPEAKER_01 25:57

What Toronto and the systems that follow in the 1960s and 70s are doing is essentially pre-computing a set of timing plans for different traffic conditions and then selecting among them based on the observed inputs. It's sophisticated compared to fixed time control, but it's still fundamentally reactive to anticipated conditions rather than the real ones, right? The city traffic engineers decided in advance what peak monitoring conditions look like and what timing plan serves it, what the midday condition looks like and what serves it, and the system switches between those plans on a schedule. Oh, it's like my phone when it shuts off at night, so nobody can wake me up in the middle of the night because all my friends live in other countries now. If the actual conditions deviate from the anticipated ones, the system doesn't adapt. Oh, like how I'm actually up at three o'clock in the morning and anybody could call me and I wouldn't care. It executes the plan that seems the most appropriate in advance. That's sort of helpful, I guess.

SPEAKER_02 26:54

I can't I like I imagine these like 60s and 70s control rooms. You know, they would look like kind of like a NASA control, you know, like Apollo, the Apollo Click. With big red buttons and stuff, you know, the LEDs, the big giant, not even LEDs, they would be the big the bulbs, yeah. On like a grid, you know, and they would light up and the big switched panels and very gray and like the map at Gettysburg.

SPEAKER_01 27:20

You ever see the map at Gettysburg or like different parts of it light? It's what it would probably Yeah, there you go, right there. Yeah.

SPEAKER_02 27:27

I'm not sure that that's what it looked like, but that's what I have in my head. Yeah, that's what it's in my head. Yeah. Uh Scoot, uh great acronym, split cycle and offset optimization technique, developed in the yeah, developed in the UK in the late 1970s and deployed through the 1980s, is the first traffic control system that adapts to live conditions. Scoot uses detector data to continuously model the flow of traffic through the network and adjusts signal timings to small and small increments every few seconds to minimize a combined objective function that uh accounts for stops, delays, and queue lengths simultaneously. Scoot continuously optimizes uh timing based on what is actually happening on the road network right now. And Scoot is the first system that treats the city's signal network as a single dynamic system to be optimized in real time.

SPEAKER_01 28:23

The progression from Scoot to modern adaptive signal control systems tracks the progression of computing power and sensor technology over the same period. What Scoot can do with the detector data and computing resources available in 1980 is generally impressive, right? What modern systems like InSync or SureTrack can do with the high-resolution video detection, distributed processing, and machine learning models trained on years of traffic data is qualitatively different. Modern adaptive systems can detect vehicle type, count vehicles in multiple lanes simultaneously, estimate cue lengths from video, and adjust timing in response to events, an ambulance approaching approaching the intersection, a school dismissal creating a pedestrian surge, a freight of train blocking, a grade crossing two intersections away. That has no pre-planned timing. That's something that no pre-planned timing scheme could anticipate. So that's I hadn't thought about that it was, I hadn't thought it was that smart. Like I didn't realize that there was that much going on. Like, okay, I think when you're going down the freeways in California, like you're on, and I'm about to say the indefinite article, you're on the 10. You know, so many people make fun of me. They're like, You're on Interstate 10. I'm like, I'm not, I'm on the 10.

SPEAKER_03 29:42

The 10.

SPEAKER_01 29:42

I'm on the 10, the 101, I'm on the five. Like, no, I'm on the, I'm on the 10. And you'll see, you can see where there's sensors, where they're, it's just tracking how fast traffic's going. Because when you go and look at it on your GPS, it'll tell you traffic's slowed, traffic's open, traffic's, you know, stopped. Well, how do you think we know that? We know that because there's cameras and there's sensors. I hadn't thought that it was everywhere, though. Like it's everywhere.

SPEAKER_02 30:04

Oh, yeah, everywhere. Well, and in the UK, I mean, it's it's lit, it's yeah, it's super everywhere. The camera structure is everywhere as as well.

SPEAKER_01 30:12

Oh, yeah, you do you do a lot of CCTV stuff over there. Yeah, yeah.

SPEAKER_02 30:15

They do tons and tons of camera. I mean, it's the most surveilled country on the, you know, on the planet. So yeah, it's you know, it's everywhere. Fortunately, you know, the surveillance, you know, systems themselves are rather, you know. Yeah. Anyway, so it's a privacy. There's a privacy discussion some other time. But yeah, like if you if you drive here in the UK, like the M25, say I use the uh the as you know for the M25, which is the big circular around the London metropolitan area, uh, and and several other motorways in the UK. They have adaptive technology for the motorways as well, so the speed limit is variable even on the motorways. So yeah, you can go 70 most of the time, and then if you don't, you know, it'll tell you, well, that's your 60, 50, 40, you know, whatever, depending upon the conditions. And that's built so that you don't get this, like everybody's going as fast as they can to then you know get to a jam, you know, a slowdown, and then everybody stops, and then basically you get that backup. So what they do is they try to slow everything down to keep everything flowing and not build up. So so that you know it's a it's a less drastic kind of ex comp expansion and contraction. It's kind of interesting. Yeah, most people. Yeah, that would never work here. Yeah, no, it wouldn't work here. They have tried it in the States. I so you know, my dad was a cop, a traffic cop, and he wrote some legislation for California and spent a lot of time on you know, traffic systems and that, but you that would and that has been tried in in different places. Didn't work. No. But but you know, the English, they like to cue, you know, they're rule followers, you know. Yeah, they are.

SPEAKER_01 32:03

So like we're just a bunch of jerks in a hurry. Like I all the time to go nowhere. Like we're just jerks in a hurry.

SPEAKER_02 32:09

California car culture, I think, would not work so well.

SPEAKER_01 32:12

No, it's all like we're offend we're we are offensive drivers, period. Yes, right. Exactly. Like you don't drive defensively in California, it's all offense. And if someone gets hit, it's because they were driving defense. Like it's you should just never do that. Just don't do that.

SPEAKER_02 32:25

I tell I tell the kids, I tell the kids don't be polite, be predictable.

SPEAKER_01 32:30

Oh, that's a good way to look at it.

SPEAKER_02 32:32

Yeah, be predictable. If you're predictable, it doesn't matter if you're polite or not. Maybe politeness, maybe politeness is the predictable thing, you know, in context, but it but don't be polite, and then that's unpredictable.

SPEAKER_01 32:45

You know, and people in the UK are you like, Oh, go ahead, go, and then somebody gets T-boned, right? Exactly.

SPEAKER_02 32:50

Like don't freaking do that. You follow the rules because that's the predictable, you know, thing. So, anyways, yeah. The the connected video dimension is where traffic signal technology is is heading next. And and we have some of that today. And it represents a more found a fundamental change than any of the sensor improvements. It's it's called vehicle to infrastructure communication. So V2I allows vehicles to communicate their presence, speed, and intended maneuver to the signal controller before they reach the intersection and allows the signal to communicate its phase state and timing to approaching vehicles. A signal controller with V2I capability knows how many vehicles are approaching from each direction, how fast they're traveling, and whether any of them are emergency vehicles requiring a preemption. A vehicle with V2I capability knows the signal's timing and can adjust the speed to arrive at the intersection during a green phase, eliminating the stop entirely.

SPEAKER_01 33:51

Okay, now the Okay, now that I know perhaps my CAPTCHA work was going towards something like this. I don't know that I'm so mad about it now. Like it doesn't seem so bad.

SPEAKER_02 34:02

Yeah, but no, it did it wasn't though. Like it wasn't Oh no, they're not. It was going towards it was going towards Waymo. It was it wasn't going towards this V2I stuff.

SPEAKER_01 34:12

No. No. Oh, all right. Damn it. They could have used it. It would have worked. That would have worked. The fuel economy, right? Like the fuel economy implications of V2I are significant enough that the Department of Transportation has cited them as a major justification for the infrastructure investment. Oh, it's infrastructure week again. The infrastructure investment. A vehicle that approaches an intersection, knowing it has 12 seconds of green remaining, can adjust its speed to arrive smoothly rather than breaking to stop and re-accelerating. Really, that's that's what we're worried about. We actually, we stopped. Like, you know what? Forget it. Like gas right now is like$6 a gallon here. So you know what? Yeah, I'm not stopping at all anymore. I'm gonna go 25 miles an hour everywhere I go, no more stopping. Eliminating unnecessarily stopped at signalized intersection could reduce fuel consumption for urban driving by a meaningful percentage across the vehicle fleet. This is the same efficiency argument that drove uh signal coordination in the 19 in 1917, applied uh with dramatically more precise information and more precise control. Well, isn't that how it always goes, though? Like we come up with an idea and then it just we just refine it and refine it and refine it. And with the more technology we have, the better we refine it until it's like crazy good and we don't even notice it anymore. Like that's it. That's just this whole podcast is about.

SPEAKER_02 35:33

Well, and I, you know, it's this sort of funny kind of the circle, you know, going around again. I think it's in the Netherlands. There's a lot of places in the Netherlands where they've gotten rid of signals because the they believe from their you know research and their studies that actually signaling messes up the flow and and in lower traffic areas it it just causes more problems than it's worth. So they just take the signals out and then let traffic do what traffic's gonna do.

SPEAKER_01 36:06

Wow. It's like their drug policy. We don't really have one. Do whatever you want.

SPEAKER_02 36:10

Do whatever you want. Do whatever you want. Which like and I it does make sense as long as it's a relatively low traffic area. But you know, when you throw in more cars, you know, if you got too many BMW drivers, uh no offense to BMW. Oh no, that yes, offense to BMW drivers.

SPEAKER_03 36:31

Wait, no, yeah. Yeah, wait, no, no, wait, yeah.

SPEAKER_02 36:35

When you throw too many people, it doesn't matter what kind of car they they've got. You throw too many of them, then you do have problems. But anyways, network traffic signals introduce a category of vulnerability that didn't exist when signals were standalone electromechanical devices. Oh, like in the Italian job when they hacked the Los Angeles signal network, right? A traffic signal network that is networked for central control and V2I communication is also networked for potential attack. Security researchers have demonstrated vulnerabilities in traffic signal systems that would allow an attacker to change signal phases, create artificial gridlock, or disable signals across a network. Most deployed traffic signal systems were designed and installed before cybersecurity was a primary design consideration. Yeah. And upgraded their security posture as a complex and ongoing project for transportation agencies.

SPEAKER_01 37:30

Well, you know, they do on on in San Francisco and other places, like on a pretty regular basis, hack the walk signs to say really awful things about Elon Musk. So for that, I say don't secure that. It makes me happy. There. I just want to put that out there.

SPEAKER_02 37:46

The uh the road, the road construction signs, you know.

SPEAKER_01 37:49

Yeah, right? Yeah, those two.

SPEAKER_02 37:51

Zombies ahead. Yeah. You know, break for dinosaurs.

SPEAKER_01 37:57

I want to go back to the governance point because I love governance, and I think it's the most unappreciated dimension of traffic signal history. Every traffic signal is making a continuous series of allocation decisions about public space and public time. How long does the car get? How long does the pedestrian get? How long, how does the cyclist get a phase at all? Does the bus get a signal priority that holds the green light while it clears the intersection? Each of these decisions reflects a value judgment about whose movement matters and whose time is worth optimizing. And those decisions are embedded in timing parameters that most people never see, and they never think the question. Although my dad listens to this podcast, and I'm pretty sure he's probably called city more than once complaining about a light. So he's got more stop signs put up in the city of Pittsburgh than any one man alive. Now, at the time that this was really successful for my dad, our mayor was named Tom Murphy, and my dad was Paul Murphy. So we just figured they they thought he was a relative, so they just kept putting up these stop signs everywhere. Like everywhere.

SPEAKER_02 39:02

It was the it was the the connection.

SPEAKER_01 39:05

Yeah, my dad was the traffic guy for Pittsburgh for like at least a couple years.

SPEAKER_02 39:10

Yeah, but I mean, my dad was a traffic cop and worked in the a traffic unit in, you know, in City Hall. And, you know, I never thought I never thought about that sort of the governance side of things, you know, where you know the single timings and progressions and things like that benefited different members of society differently. You know, it just like like it's because someone had to think about that.

SPEAKER_01 39:37

Yeah, someone had to sit down and think about that, right? Yeah. It's just does that corner have a ramp? If it does, does that mean a person in a wheelchair is gonna go across it? If they do, how long does that take? Is that faster or slower? Like, yeah, you have to really think about that. It's a real crazy.

SPEAKER_02 39:52

It's something that I just hadn't really thought about. And I love like I learned my dad taught me how to how to how to do direct traffic, you know, with the whistle and the gloves and the scouts, we had to do that once in a while. And so, you know, I I I like traffic signals and I like, you know, it's all yeah, it hits all the nostalgia buttons, but yeah.

unknown 40:13

Yeah.

SPEAKER_02 40:15

Uh traffic signal priority, which allows buses and light rail vehicles to extend uh a green or advance a red when they're uh behind schedule is one of the cleanest and clearest examples of a signal timing decision that embodies policy choice. A city that implements, I always wonder how buses in the UK get you know places on time. And they change the lights. You know, they change the lights. I I don't know if that's the case, you know, in most of where, you know, I'm in in the UK, but like the buses run they pretty much on time. A city that implements transit signal it's transit signal priority is explicitly deciding that the reliability of bus service for the people who depend on it is worth a small delay to the cars that would otherwise get that green phase. The cities that have implemented see measurable improvements in transit schedule adherence and ridership. Cities that haven't implemented it's implicitly decided that car uh throughput takes precedence over bus reliability.

SPEAKER_01 41:17

The pedestrian scramble phase, where all the vehicle movement stops simultaneously and the pedestrians can cross in any direction, including diagonally, is another example of signal configuration that embeds a choice about who the intersection is for. The scramble phase is slower for vehicles because it requires a longer total cycle to give pedestrians their time. It's significantly safer and more convenient for the pedestrians, especially at busy urban intersections where diagonal crossing is a natural desire line. Cities have installed scramble phases, most prominently in Tokyo, because that's when when you say that to me like that, like you can cross, I mean there's something like that's the big one.

SPEAKER_02 41:59

Yeah.

SPEAKER_01 41:59

Tokyo. Like everything you ever see is that crosswalk, right? And increasingly, various North American cities have made the judgment that pedestrian convenience and safety justifies the the vehicle throughput cost. The choice is available. Most cities haven't made it. I don't think I don't remember there being that in LA anywhere. I don't remember because it's not a really it's not a pedestrian town.

SPEAKER_02 42:23

There might be there might be one in downtown. There might be one in downtown. Like over by the staple center. Oh, those are big intersections. Vaguely remember that, but yeah, I I don't I don't have to look at an overhead. Because in in uh from what I remember in LA, the if there is a diagonal, it there's a diagonal lines on the uh in the intersection. Yeah, yeah, yeah, yeah. Here in the UK, the one I know is uh in London near in Covent Garden. So what is it, uh Charing Cross Roads? Um right around not too far from Leicester Square, there's one right by the theaters and stuff like that. The the automation of signal timing through adaptive systems has an interesting effect on the governance question because it displaces the decision making from human engineers making explicit choices to optimization algorithms pursuing defined objective functions. If the objective function prioritizes vehicle throughput, the system will optimize for vehicle throughput. Duh. And the pedestrian crossing time will be what's left over after the vehicles are served. If the objective function weighs pedestrian delay and vehicle delay equally, the system produces a different timing. The values sit in that objective function. The objective function is a political document written by engineers who may or may not have been thinking about it in those terms.

SPEAKER_01 43:49

There's a version of the smart city vision where adaptive traffic signals combined with connected vehicles and real time transit data create an urban mobility system that moves everything. Everyone efficiently without any individual waiting unnecessarily. Aw. It's a compelling vision. It also requires every participant in the system to be visible to the system, to be carrying a device that communicates with the infrastructure. The pedestrian without a smartphone, the cyclist without a transponder, the driver with an older vehicle without V2I capabilities are invisible to the system the same way the cyclist is invisible to the inductive loop today. The efficiency gains of a fully connected system accrue to the connected participants. The unconnected remain at the residual and they're probably super angry that they can't get anywhere because everybody else is kind of taking up their space. That would be so irritating. I can't even tell you.

SPEAKER_02 44:44

But but don't you think that with autonomous systems, autonomous driving systems, that like a traffic engineer's dream is no humans making decisions. Right? Right. And the vehicles coordinating with the system to out, you know, to produce the totally optimal systems, right? Traffic flowing constantly, pedestrians and cyclists getting their way, buses being on time, trams and tr and trains and things, you know, moving appropriately. Like that only works if no human gets to make a decision, you know?

SPEAKER_01 45:24

Right. Yeah, and everybody gives up to that. Like I agree to be surveilled in order for that to happen.

SPEAKER_02 45:29

That's right. Not just surveilled, but but my I give up my my agency to to the system to allow an efficient outcome. And I might give up my like I might have been able to get to work five minutes faster, you know? If I just did it my way. But if everybody did it just their way, then then the system collapses and nobody gets to get to work five minutes earlier, you know? And it's just I don't know. I think especially in in the States, that it just yeah, I don't think it flies.

SPEAKER_01 46:08

It wouldn't work. No, it wouldn't work. We're too we are too individualistic to pull this off. Like that we all stop for red lights is about all we can handle. I think. Like I think that's enough. We shouldn't ask for the colour.

SPEAKER_02 46:19

The green arrow, the green arrow.

SPEAKER_01 46:20

Everybody rules everybody everybody follows the green arrow. Yeah. Yeah. Like I I think that's all you can ask of us, and I think that's all we could probably do. I mean, I've met some of us. I think that's all we could probably do.

SPEAKER_02 46:31

I think in maybe like like perhaps in p some places in China or Singapore, like even Singapore though, there's got some yeah, there's some interesting car dynamics in Singapore, but I think China should be able to pull it off, right?

SPEAKER_01 46:47

They they have autonomous vehicles. Those they seem to work. They have straight streets because that seems to work. Like, you know, like there's kind of set up for that. There's not I mean, there at no point is there a Chinese city that looks a lot like Boston, right? Like there's they just don't exist. And so they might be able to pull it off, right? Like come up with that big or and and they're already being surveilled. They it's a culture that they don't have a problem with. So and they participate in those kinds of things. Like, hey, China, like let's call them, tell them, hey, you want to try this? Because we want to see if it works, if it actually works.

SPEAKER_02 47:20

Yeah, I didn't do a ton of research on on traffic systems in China, but I have I have seen some, you know, some of those systems, and they are you know more advanced than what you see in you know Europe or or or the states, and that's simply because the infrastructure is built from the ground up with a more, let's say, holistic of view of the optimizing the objective.

SPEAKER_00 47:42

Oh, that was a nice way to say it. No, that was nice. Yeah.

SPEAKER_02 47:45

But but I I yeah, I think eventually it has to come. You know, that it's it will eventually get here. Maybe it's not 10 years. But it's 50 years.

SPEAKER_01 47:54

Yeah, yeah. I'll be dead by then. I don't really care what it's like. And I'll be able to do road trips all the way up until then. So yay! Yeah, ridiculous to drive.

SPEAKER_02 48:03

You know, in in s in places like okay, so I don't know if you uh you know if you've been to you've been to Madrid, right? Um Yeah, so Madrid used to be a place where you drove through the drove, it was a driving, you know, you could drive in the city, but now in the middle, you can't drive at all. Like there are no cars. And that was that's like that's a that was a kind of a shock to the system for a lot of people. Paris is doing a similar thing where they're you know, and London has done that as well, that where they're reducing the places where cars can be. And to me, to me, like I f just having been experienced Madrid in both before and after, you know, and experienced, you know, London and Paris again before and after some of these changes. I I like it after, you know. I like it when there are less cars and public transportation works better. And I think that that's like almost the first step, you know, because if you follow that first step, the next step is well, not just fewer cars, but more control about how the cars operate in the spaces that they're allowed to be.

SPEAKER_01 49:12

So So that was like Santa Monica's plan for a smart city was you did not, you did not have a car. You you're you would you your car would drop you off somewhere where you are it was allowed to drop you because it's assuming it has autonomous fleet, right? So your car would drop you off, it would go to one of the six parking structures in in Santa Monica, it would park itself and you would just walk everywhere. You would go, it would come pick you someplace else. You would go and it was a place that was designed for things to pick you up. And then that way people weren't in the streets, people weren't just stopping and pulling over, people there weren't huge lines trying to get into the parking garages. You didn't have to look up whether the parking garage was ready or not, because you were never gonna be part of that anyway. Like all of that, like like, yeah, I mean, like I think any smart city thinks, why the hell would we have a car here? Like, I think they all think like that. I think they all think, and that's not a bad way to think, right? Like, I think that probably makes sense. But on the other hand, are you gonna have a tram? Like, how are people gonna get around? Like, especially if you're a larger, like downtown area or whatever. Like, you gotta think about that too. But you know, that's how the smart city designs are thinking. Yeah.

SPEAKER_02 50:15

Well, Santa Monica is not exactly set up to be so smart, you know? I like there's like it's kind of spread out, you know.

SPEAKER_01 50:24

There's this There's a light rail that takes you straight from downtown to Santa Monica now, though.

SPEAKER_02 50:28

Yeah, there is. It's true. Yeah, it's true.

SPEAKER_03 50:31

Yeah. Pretty great, actually.

SPEAKER_02 50:32

I mean, but uh but also th think about you know the politics that are involved to actually make that happen, right? Oh yeah, I I can't decades, decades of resistance, right? You know, and unfortunately LA LA had a great public transit with you know the the rails, the the you know, the uh the red line, and it got yanked out, right, because of lobbying and and you know, and it's a drive in town, dude.

SPEAKER_01 50:59

It's a drive-in town, yeah. Yeah, but but it wasn't like five friends would take five different cars to go to the same place. Like one person wouldn't drive around and pick everybody up. Everybody brought their own cars. Yeah. Yeah, yeah, yeah.

SPEAKER_02 51:11

I mean that's I that's how it was when I grew up as well. But yeah, but see, it could have it could have been better. It could have been better.

SPEAKER_01 51:19

It could have been better. Well, most things. Like it's why we can't have nice things. I'm stupid, you know? Yeah.

SPEAKER_02 51:24

You can I think the some of the lines for the red car are still around, but anyways. The the traffic lights history is a clean example of a technology that seems simple. The complexity is real, and it's mostly hidden from the people who interact with it. The light you stop at is visible, is the visible surface of a system that includes includes detector hardware, controller software, sensors, communication infrastructure, central management platforms, and a set of timing parameters that encode decades of engineering judgment and political choice. None of that is visible from the driver's seat. Well, the lights are visible, right? You see the red and you stop. You see the green and you go. And you see the yellow and you mash the gas. Go faster. The compliance is so automatic that most people have never consciously thought about why they do it.

SPEAKER_01 52:11

The railroad origin is the thing I keep coming back to because it's such a clear example of how new technologies borrow their interface conventions from the technologies they replaced or extended. The traffic light is red and green because trains are red and green because maritime signal flags are red and green, and the whole chain of inheritance is invisible to the person stopping at the light. The convention feels natural and inevitable, and it's actually a series of historical accidents that happened to produce something durable. Red means stop because someone on a dock 150 years ago chose a red flag to mean danger. And that choice propagated through railroad signals into the very urban intersection, into every urban intersection on planet Earth. Oh, and probably in Artemis 2. I'm pretty sure like the red buttons in there mean bad.

SPEAKER_02 53:03

Yeah, definitely. The the adaptive system trajectory is probably the most important development in traffic signal technology since the three-color phase. It it is the first time the signal network can respond to live conditions instead of to predictions made in advanced. A fixed time signal is a bet about what traffic will look like. An adaptive signal is a response to what traffic actually looks like. The efficiency difference, oh, and and now we want to have predictive models, right? The efficiency difference compounds as sensor resolution improves as connected vehicles provide more granular data and as the optimization algorithms get better. The intersection you stop at in 2035 will be m will be making better decisions and and predictions in real time than the intersection you stop at today. Most drivers will never notice a difference.

SPEAKER_01 53:53

The most obeyed command in human history is a colored light that was installed by a police officer at a London intersection in 1868 and then promptly exploded. Oh, the light, not the officer. Everything since then has been I love misplaced modifiers. Everything since then has been incremental improvement on that original idea, which is that if you give people a clear, consistent, shared signal about whose turn it is, they will mostly follow it. Not because they have to, because the alternative is chaos. And most people, most of the time, prefer not to live in chaos, right? The traffic light works because it solved the coordination problem and then taught everyone the solution simultaneously at every corner for 150 years until following it became a reflex. That's a technology succeeding on a civilizational scale. And it happens so quietly that you don't notice it when the light is oh, you only notice it when the light is broken, and suddenly nobody knows what to do. Like that is always like everybody like you go. Like, but why am I still? I'll go. Okay, okay, you go wave on the hands. Yeah, like I'm fine, I'm going. I'm going. This is me every time the light's broken. Yeah. It causes me existential dread. I see it blinking and I'm like, oh dear god, can I turn left here? Like I don't want to do that.

SPEAKER_02 55:14

Then they'll like the inching out, you know, and then yeah, yeah. Yeah. I I think I think there are definitely some cities that prefer chaos, though. Like, like I said, Cairo, man. Cairo definitely loves chaos.

SPEAKER_01 55:28

All those Asian countries, too.

SPEAKER_02 55:29

Like it'll I always want to like uh you know Mumbai.

SPEAKER_01 55:33

Oh yeah, places like that, you're like, oh, how do you guys know what you're doing? Like, how do you know? I don't understand. Yeah, that's funny.

SPEAKER_02 55:40

Yeah. If you've ever sat at a red light at 2 a.m. with no traffic in any direction and felt the specific existential weight of that moment, you're grappling with the gap between the rule and its purpose. And have you ever have you ever like sat there for long enough that like you realize you ride the O.

SPEAKER_01 56:02

You ride the O. The O thinks there's something there. That's what riding the O's are all about.

SPEAKER_02 56:07

I I've I have backed up to hit the O to then go, you know, to see.

SPEAKER_01 56:12

Yeah, to turn the light on.

SPEAKER_02 56:14

Yeah, I I I will admit that once I I it wouldn't change, so I had to go.

SPEAKER_01 56:20

Yeah, so you just go. There's no one else there. Who's gonna see you? If you get pulled over, I would tell the cop how long did you want me to sit here? Yeah, till tomorrow? Yeah.

SPEAKER_02 56:27

Uh you you've been grappling with the gap between the rule and its purpose, and philosophers have been writing about the gap for centuries.

SPEAKER_01 56:35

If you've ever been in a city with a pedestrian scramble phase and the first for the first time, and you felt an irrational joy of being able to walk diagonally across an intersection, that joy is legitimate. And the cities that haven't installed scrambled phases should think about why.

SPEAKER_02 56:53

I love scramble phases. And if you've ever watched a traffic engineer's eyes light up when you ask them about signal coordination, please let them talk. They have things to say and almost no one asks.

SPEAKER_01 57:11

Yeah, we do. This is the nostalgic nerds where the infrastructure you stop at twice a day turns out to contain 150 years of engineering politics and gas explosion in Westminster. Thanks everyone for tuning in. Please subscribe, follow, and share with someone who has opinions about yellow flight duration. They exist at every intersection, and we'll be back next week. Mark, thank you.

SPEAKER_02 57:33

Thank you very much.

SPEAKER_04 57:53

In between, I gotta dive. I don't get time in the scene. Before I begin, I'm gonna be like I've never been.