Smart EV Routing System Minimizes Traffic Accident Impact
In a groundbreaking development aimed at transforming the electric vehicle (EV) user experience, a team of researchers from Chongqing University has introduced a revolutionary path planning and charging navigation strategy designed to mitigate the disruptive effects of traffic accidents. As EV adoption accelerates globally, the challenges of urban congestion and unforeseen incidents like collisions have become critical pain points for drivers. The new system, detailed in a study published in the prestigious journal Power System Protection and Control, leverages real-time traffic data and a sophisticated dynamic model of accident consequences to deliver an intelligent navigation solution that significantly reduces travel costs and improves road efficiency.
The research, led by Huang Bo, Hu Bo, Xie Kaigui, Shao Changzheng, Lin Chengrong, and Huang Wei from the National Key Laboratory of Power Transmission Equipment Technology at Chongqing University, addresses a significant gap in current navigation technologies. While existing systems are adept at handling predictable, recurring traffic congestion—such as rush-hour jams— they are largely ineffective against the chaos caused by random, unpredictable events like traffic accidents. These “sporadic” incidents, as the authors term them, can cripple road networks, leading to extended delays, increased energy consumption, and heightened anxiety for EV drivers concerned about running out of power.
The core innovation of this new strategy lies in its ability to accurately quantify the dynamic impact of a traffic accident from the moment it occurs until the congestion fully dissipates. Traditional navigation models often rely on simple “flow-time” relationships, which assume that travel time increases linearly with traffic volume. However, this approach fails in the context of a major accident. When a collision blocks lanes, the actual traffic flow on that stretch of road can plummet to near zero, while vehicles queue up for miles, resulting in travel times that are vastly longer than what conventional models can predict. This inaccuracy can mislead drivers, causing them to enter a congested zone and become part of the problem.
To overcome this limitation, the Chongqing University team developed a novel consequence assessment model that captures the entire lifecycle of a traffic incident: its “occurrence-sustenance-dispersion” process. This model begins by evaluating the severity of the accident, specifically how many lanes are blocked, which directly impacts the road’s remaining capacity. Using data from the Highway Capacity Manual, the model calculates this reduced capacity, providing a more realistic foundation for its predictions.
The model then simulates the evolution of the traffic queue. It considers the initial moment of the accident, the time it takes for traffic monitoring centers to detect it, the arrival of emergency crews, the duration of the cleanup, and finally, the time required for the residual congestion to clear. By integrating real-time traffic flow data into this dynamic framework, the system can provide an accurate and continuously updated estimate of the delay on the affected road. This is a crucial advancement, as the severity of a traffic jam can change minute by minute, and drivers need the most current information to make sound decisions.
The heart of the strategy is a comprehensive road impedance model. Rather than using a simple metric like distance, this model calculates a “composite impedance” for every road segment, combining two factors that are paramount to EV drivers: travel time and energy consumption. These two metrics are converted into a unified cost, weighted by the user’s value of time and the cost of electricity. This allows the system to present a holistic view of a route’s cost, balancing the desire for a quick journey against the need to conserve battery power.
This composite impedance is not static. It is recalculated in real time based on the dynamic traffic conditions, including the precise delays predicted by the accident consequence model. The result is a constantly updating map of the road network, where the “cost” of each path reflects its true, current state.
To turn this real-time data into actionable navigation advice, the researchers devised a rolling optimization algorithm based on the classic Dijkstra’s algorithm. Unlike traditional static pathfinding, which calculates a single route at the start of a journey, this rolling approach is dynamic. As an EV travels, the system continuously re-evaluates the network. If a new accident is reported or traffic conditions change significantly, the algorithm instantly recalculates the optimal path based on the latest composite impedance matrix. This ensures that the driver is always being guided along the most efficient route available at that precise moment, dynamically rerouting to avoid emerging congestion.
The practical benefits of this system were rigorously tested using a coupled model of a 12-node traffic network and the IEEE 33-node power distribution grid. The results were compelling. In a simulation where a traffic accident occurred on a major road, the new real-time navigation strategy outperformed conventional static shortest-path routing by a wide margin. For a typical journey, the total cost—including time and energy—was reduced by over 90%. Drivers using the static system were funneled into the accident zone, where they faced long queues and high costs, while those using the dynamic system were seamlessly rerouted onto alternative paths, avoiding the worst of the congestion.
The analysis of vehicle counts on the accident-affected road provided a stark visual of the system’s effectiveness. Under static navigation, the number of vehicles on the blocked road surged, creating a severe and prolonged traffic jam that took over an hour to clear. In contrast, the real-time strategy successfully diverted traffic, keeping the vehicle count on the accident road relatively stable and preventing a catastrophic buildup of congestion. This not only benefits the individual driver but also enhances the overall efficiency of the entire transportation network. The study quantified this by showing that the average road travel efficiency was improved by over 180% in the first 30 minutes after an accident and by more than 300% in the subsequent 30 minutes compared to the static approach.
The system’s impact extends beyond just the roads; it also positively influences the power grid. Traffic accidents can cause a ripple effect on charging patterns. When drivers are stuck in traffic, their arrival at charging stations is delayed, leading to a sudden, concentrated influx of vehicles once the congestion clears. This creates sharp peaks in electricity demand, which can strain the local distribution network. The study demonstrated that the real-time navigation strategy smooths out this charging load. By preventing massive queues and distributing traffic more evenly, it ensures that EVs arrive at charging stations in a more staggered manner. This reduces the maximum peak load on the grid by a significant margin—up to 600 kW in the simulation—and also shortens the duration of the elevated load by nearly 30 minutes. This stabilization of the charging demand is a critical benefit for grid operators, helping to maintain voltage stability and reduce the risk of overloads.
The research also delved into the decision-making process of EV drivers, particularly when it comes to charging. The system incorporates a sophisticated charging navigation model that factors in not just the cost of electricity, but also the time spent waiting in line at a busy charging station. Using a queuing theory model (specifically, an M/M/c model), it can predict the average waiting time at a charging station based on the current arrival rate of vehicles and the service rate of the chargers. This information is then integrated into the overall cost calculation. A driver is not just being told to go to the nearest or cheapest charger; they are being guided to the one that offers the best combination of low energy cost, short travel time, and minimal waiting time.
A sensitivity analysis revealed the robustness of the strategy. The researchers tested how changes in the driver’s “value of time”—a parameter that reflects how much a user is willing to pay to save an hour of travel—affect the navigation outcome. For a wide range of reasonable values (from 6 to 18 yuan per hour), the system consistently recommended the same optimal path, which avoided the accident zone. Only when the value of time was unrealistically low did the system suggest a path through the congested area, confirming that for most users, the time savings from avoiding an accident far outweigh the potential extra distance.
To prove the scalability of their approach, the team conducted a final test on a much larger, 35-node traffic network with two simultaneous accidents. The results confirmed that the strategy remained effective, successfully guiding vehicles around multiple problem areas and significantly reducing the total travel cost for users compared to static or less sophisticated dynamic methods. This demonstrates that the core algorithm is not limited to small, controlled environments but has the potential to be deployed in complex, real-world metropolitan areas.
The implications of this research are profound. It represents a major step forward in creating truly intelligent transportation systems that are resilient to random disruptions. As cities worldwide strive to meet carbon reduction targets by promoting EVs, the infrastructure to support them must evolve. A navigation system that can anticipate and mitigate the effects of traffic accidents is no longer a luxury but a necessity. It enhances user satisfaction by reducing stress and saving time and money, promotes more efficient use of existing road infrastructure, and supports the stability of the electrical grid by smoothing out demand. The work of Huang Bo and his colleagues at Chongqing University provides a powerful blueprint for the next generation of smart mobility, where vehicles are not just electric, but also intelligent, adaptive, and connected in a way that makes our cities safer, cleaner, and more efficient for everyone.
Huang Bo, Hu Bo, Xie Kaigui, Shao Changzheng, Lin Chengrong, Huang Wei, National Key Laboratory of Power Transmission Equipment Technology(Chongqing University), Power System Protection and Control, DOI: 10.19783/j.cnki.pspc.240089