Introduction
As the adoption of electric vehicles (EVs) accelerates, the demand for charging infrastructure has surged. Traditional grid-to-vehicle (G2V) charging stations often suffer from heavy overloads and long queues, creating bottlenecks in urban transportation. Vehicle-to-vehicle (V2V) charging—where an EV with surplus energy (EVD) provides power to an EV in need (EVC)—presents a flexible, decentralized alternative. However, coordinating these mobile energy exchanges requires solving complex problems related to communication overhead, navigation efficiency, and optimal matching.
The Intelligent Direct V2V Charging Navigation (IDCN) Framework
This paper introduces the Intelligent Direct V2V Charging Navigation (IDCN) strategy, a semi-centralized framework designed to optimize the coordination of V2V charging. The system leverages several advanced technologies to ensure scalability and efficiency:
- MEC-Based Architecture: By utilizing Mobile Edge Computing (MEC) servers attached to Roadside Units (RSUs) or public transit infrastructure, the system offloads data mining and aggregation. This reduces latency and prevents the communication bottlenecks typical of fully centralized models.
- LSTM Traffic Prediction: A Long Short-Term Memory (LSTM) neural network is employed to predict traffic flow and density, providing accurate travel time estimations for EVs.
- Q-Learning for Route Selection: An adaptive Q-learning algorithm assists EVs in selecting optimal paths to meeting points, balancing travel time against energy consumption.
V2V Matching and Navigation Strategy
The core of the IDCN strategy lies in its approach to selecting stopping locations (SLs) and pairing vehicles. Unlike "greedy" models that simply direct vehicles to the nearest available charger, IDCN adopts a global trip perspective. It evaluates SLs based on the vehicle's entire trajectory to its final destination, minimizing the "detouring cost" in terms of both time and energy.
The pairing process is formulated as a maximum weighted matching problem on a bipartite graph. The Kuhn-Munkres (KM) algorithm is used to find the global optimum for pairings, considering factors such as energy gaps, arrival times, and the total amount of energy to be exchanged. To prevent "synchronous conflicts"—where multiple vehicles converge on the same "best" location—the authors introduced a "charging comfortable degree" metric to distribute demand more evenly.
Experimental Evaluation
The IDCN strategy was tested using a realistic urban simulation of a 4 km × 3.7 km area in Nanjing, China. The researchers utilized a practical dataset of mobile vehicle movements provided by the Nanjing Transport Bureau (January–April 2018). Vehicles were modeled after the Tesla Model S (85 kWh battery) to ensure realistic energy consumption and charging rates.
Key Results
The proposed IDCN strategy outperformed several benchmarks (including STCC, DVCS, and GCNA) across multiple metrics:
- Efficiency: IDCN achieved the lowest mean detouring travel time (~89.6 seconds) and the lowest detouring energy consumption (~0.32 kWh) per vehicle.
- Wait Times: Precise arrival predictions and reservation queue management significantly reduced average waiting times at SLs.
- Scalability: The MEC-based approach kept transmission costs stable even as the number of EVs increased, maintaining a high packet delivery ratio.
- Matching Success: The KM-based algorithm increased both the total energy exchanged and the number of successfully charged vehicles compared to simpler first-come, first-served rules.
Conclusion
The IDCN framework demonstrates that intelligent navigation for V2V charging can significantly alleviate the pressure on traditional charging grids. By combining edge intelligence with a global trajectory perspective and advanced matching algorithms, the system provides a scalable and efficient solution for the future of mobile electric vehicle energy exchange. Future work will focus on adaptive offloading schemes and incorporating individual driver behavior into the navigation models.
Summary updated