Getting your Trinity Audio player ready...
|
The need for efficient and sustainable transportation solutions has never been greater in a rapidly urbanising world. Researchers at Berkeley are at the forefront of this effort, pioneering a revolutionary traffic signal control algorithm, HumanLight, which uses reinforcement learning to prioritise high-occupancy vehicles (HOVs) at intersections. This cutting-edge technology promises to transform urban mobility by encouraging environmentally friendly transit options over single-occupancy cars, significantly reducing travel times and promoting sustainable transportation choices.
HumanLight is the brainchild of Dimitris Vlachogiannis, Scott Moura, Jane Macfarlane, and Hua Wei, whose work is detailed in Transportation Research Part C. This innovative approach to traffic management represents a significant leap forward, offering a more democratic and sustainable solution to urban congestion.
Imagine a scenario where catching a ride on a high-occupancy vehicle is more environmentally friendly and faster than driving alone. This is the vision shared by the Berkeley research team. By leveraging a simulated environment, the researchers have developed and tested HumanLight, a traffic signal control algorithm designed to maximise the throughput of people rather than vehicles at intersections. The technology uses reinforcement learning, a form of artificial intelligence, to prioritise and reward passengers of HOVs with more green lights.
The study’s lead author, Dimitris Vlachogiannis, and Co-authors Scott Moura, Jane Macfarlane, and Hua Wei contributed significantly to this research. Scott Moura, the Clare and Hsieh Wen Shen Professor in Civil and Environmental Engineering, and Jane Macfarlane, Director of the Smart Cities Research Centre, have discussed their work’s implications with Berkeley Engineering. Their insights provide a deeper understanding of how HumanLight could someday offer a more democratic and sustainable traffic management solution.
Jane Macfarlane’s experience, which used seat sensors to detect the number of occupants in a vehicle during emergencies, inspired the idea of HumanLight. This technology could similarly be used to develop a traffic signal control system that prioritises vehicles with higher occupancy.
Scott Moura emphasises the importance of traffic management in addressing global greenhouse gas emissions. With transportation accounting for 40% of California’s greenhouse gas emissions, innovative solutions like HumanLight are crucial for achieving climate goals. While widespread adoption of electric vehicles is one approach, optimising existing infrastructure through intelligent traffic management offers a more immediate and cost-effective solution.
HumanLight represents the next step in the evolution of transit signal priority and emergency vehicle preemption. Unlike traditional control systems, HumanLight uses reinforcement learning to manage the dynamic behaviour of complex traffic environments. This approach allows for more intelligent traffic light timing, prioritising the flow of vehicles based on occupancy. For example, at a four-way intersection, HumanLight can adjust the timing to prioritise directions with higher occupancy vehicles, ultimately moving more people more efficiently.
The findings from the study revealed that HumanLight not only benefits high-occupancy vehicles but also ensures that those in single-occupancy vehicles are not significantly disadvantaged. This democratic solution encourages carpooling and public transit, reducing overall congestion and emissions. The research highlights the potential of HumanLight to incentivise sustainable transportation behaviours while improving traffic flow.
Before HumanLight can be widely implemented, several steps need to be taken. Aligning stakeholders, including infrastructure owners and city authorities, is crucial. Additionally, the necessary infrastructure must be in place, such as traffic cabinets equipped with radios to communicate with vehicles and the cloud. The U.S. Department of Transportation’s Vehicle-to-Everything (V2X) deployment plan aims to equip every intersection in the United States with this technology by 2035, paving the way for innovations like HumanLight.
HumanLight represents a significant advancement in traffic signal control, offering a sustainable and efficient solution to urban congestion. Innovations like HumanLight will be essential in creating more sustainable and livable urban environments as cities grow.