The road to fully autonomous vehicles
By Jack Boreczky and Nicholas Chan
Imagine an interconnected world with streets free of traffic signals: where vehicles operate without human control; where pedestrians cross without worrying about a collision; and where vehicles communicate with infrastructure to eliminate traffic jams.
Sounds great right?
Well, here’s the good news: What was once considered a utopian fantasy may soon become reality.
Motivated by a desire to improve traffic conditions and pave the way for fully autonomous vehicles, Dr. Kun Zhou of the Institute of Transportation Studies at the University of California, Berkeley is leading the development of the Multi-Modal Intelligent Traffic Signal System (MMITSS, pronounced “mits”), and the connected vehicle testbed, which work in tandem to minimize accidents on the roadway and eliminate traffic jams once and for all. According to Zhou, implementing the MMITSS will mean that” in the future, we [will] not need any traffic lights at all.”
A set of over 30 new traffic signal components that can be seamlessly integrated as additions to our existing traffic control systems, the MMITSS is the infrastructure foundation that determines efficient traffic light patterns. Connected vehicles, on the other hand, are vehicles equipped with an array of on-board sensors such as cameras, geographical positioning (GPS), and light detection and ranging (LIDAR), that all help the real-time traffic conditions to the MMITSS.
While traditional traffic signals merely flicker in their cyclical red-yellow-green pattern and use loop detectors — magnetic sensors embedded at the start of intersections — to make small adjustments based on the number of vehicles physically passing over the intersection’s loop detector (the car’s metal body triggers the magnetic loop detector), the MMITSS completely changes the approach.
Unlike loop detectors in traditional traffic light systems, the MMITSS combines data from connected vehicles and sensors mounted atop each traffic light pole to measure the actual number of cars in each branch of an intersection. Instead of estimating volume by the number of cars that pass over the magnetic sensor in one green light period, this new approach can help determine optimal changes in light colors.
In other words, these lights can prioritize branches of an intersection that have more cars waiting by providing a longer green light period for those branches. Each traffic light also relays this data with others in the network, allowing the algorithm to predict required signal changes at further intersections based on nearby traffic patterns. Instead of rolling up to a red light and waiting for an eternity, imagine if the traffic light would turn green right as you approach!
Zhou uses an example to explain this: in a line of ten vehicles following one another at a normal intersection, each driver must independently predict if they can pass through the green light as they approach the signal. However, the line may be partially cut by a red light, forcing the following cars to stop, which can become a catalyst for congestion.
At an MMITSS-equipped intersection, however, the traffic light would know of the impending approach of the ten cars, treating it as a single “unit” and recognize that “[since] there are 10 cars approaching, I’ll create a single window [to allow them] all to pass.”
This traffic signal then relays this data to all ten vehicles, allowing all ten cars to pass the intersection together, rather than forcing each driver to guess the remaining “green time” or executing dangerous maneuvers to cross before the light turns red. The network of MMITSS-equipped traffic lights then feeds data to the other lights, allowing them to make predictive determinations on the optimal light timings. This makes the intersection more efficient and safer — as well as less frustrating.
Aside from the safety and experience improvements, the MMITSS also provides long-term cost reductions for municipalities, stemming from the reduced maintenance and upkeep costs of traffic signals. While the MMITSS can cost up to $10,000 to install at each intersection, Zhou says that long-term benefits can be realized rather quickly. Since the MMITSS allows maintenance staff to monitor the status of every light from a single dashboard, it “reduces the human effort to maintain [traffic signals],” as maintenance crews staff no longer have to physically inspect each light at each intersection. As a side effect of increased connectedness, signals will be able to report issues and trigger maintenance requests themselves.
Through its first phase of testing, researchers have found that the systems significantly increase the efficiency of intersections. The main testing area for the project is a series of 11 intersections in Palo Alto known as the Connected Vehicle Test Bed. Zhou discussed a test in which connected trucks were sent through these intersections and their travel time measured. Even with only a few connected vehicles (the testbed is on a busy public intersection with both connected vehicles and traditional vehicles), the traffic efficiency increased by nearly ten percent. However, while these tests proved the effectiveness of the MMITSS, guaranteeing safety and reliability is a major hurdle before the systems can be implemented on a wider scale. While Zhou says this is a challenging task, he strongly believes that the MMITSS will undoubtedly reduce the number of accidents on roadways.
However, improved infrastructure is not the only component of Zhou’s research. The array of sensors on connected vehicles not only pass data to infrastructure, but also prevent accidents by sensing hazards independently. The most notable safety features include emergency braking, which can reliably stop vehicles in the face of unexpected obstacles, such as when an animal or jaywalker suddenly appears in front of the vehicle. Working in tandem with MMITSS, this braking signal can then be passed to cars behind it, eliminating the high-speed rear-end collisions that often result from emergency braking.
Ultimately, Zhou believes that the MMITSS and connected vehicles will be used in tandem to perfect the development of autonomous vehicles. He believes that “transportation will be different with connectivity.” By communicating with the infrastructure and other vehicles, self-driving cars will become much safer and more reliable. In fact, this research may eventually prove to be the missing piece to the autonomous vehicle puzzle. Eventually, he envisions that roadways will be populated by vehicles that drive safely, without the need for traffic signals or even human control.
The work that engineers do shapes the world around us. But given the technical nature of that work, non-engineers may not always realize the impact and reach of engineering research. In E185: The Art of STEM Communication, students learn about and practice written and verbal communication skills that can bring the world of engineering to a broader audience. They spend the semester researching projects within the College of Engineering, interviewing professors and graduate students, and ultimately writing about and presenting that work for a general audience. This piece is one of the outcomes of the Fall 2019 E185 course.
