Smart Braking: New EV Strategy Boosts Energy Recovery
In the rapidly evolving world of electric mobility, every kilowatt-hour counts. As automakers race to extend driving range and improve efficiency, one of the most promising frontiers lies in the intelligent management of energy during deceleration. A groundbreaking study from Jiangsu University has introduced a sophisticated new regenerative braking control strategy that could redefine how electric vehicles (EVs) reclaim energy on the move. By integrating real-time road surface identification with a nuanced understanding of dual-motor dynamics, the research team has achieved a remarkable 65.55% braking energy recovery rate under standard conditions—setting a new benchmark for efficiency in four-wheel-drive EVs.
The study, led by Professor Pan Gongyu and graduate researcher Xu Shen from the School of Automotive and Traffic Engineering at Jiangsu University, was published in the January 2024 issue of the Journal of Jiangsu University (Natural Science Edition). Their work addresses a fundamental challenge in EV design: how to maximize energy recovery without compromising safety or stability. While regenerative braking is a standard feature in modern EVs, its effectiveness is often limited by conservative control strategies that fail to fully exploit the available traction between the tires and the road. Traditional systems typically assume a fixed road condition, leading to suboptimal performance, especially on variable or low-grip surfaces.
Pan and Xu’s approach breaks from convention by treating the road not as a static backdrop but as a dynamic variable that must be continuously assessed. At the heart of their strategy is a novel road identifier that estimates the peak adhesion coefficient—the maximum grip available—between each tire and the road surface. This isn’t a one-size-fits-all measurement; instead, it’s a per-wheel calculation that accounts for differences in load, wear, and local road conditions. The system compares real-time data on wheel slip and road adhesion against a library of eight standard road types, ranging from dry asphalt to icy pavement. Using fuzzy logic, it determines the similarity between current conditions and each reference surface, then computes a weighted average to arrive at a precise estimate of the peak adhesion coefficient.
This level of granularity is critical. In conventional systems, an inaccurate estimate of road grip can lead to two equally undesirable outcomes: either too little regenerative braking, wasting recoverable energy, or too much, risking wheel lockup and loss of control. By dynamically adapting to actual road conditions, the Jiangsu University team’s strategy ensures that the vehicle operates as close as possible to the optimal slip ratio—the point at which friction and adhesion are balanced for maximum energy recovery without sacrificing stability.
But sensing the road is only half the equation. The other half lies in how the vehicle allocates braking force between its electric motors and hydraulic friction brakes. The researchers focused on dual-motor, four-wheel-drive EVs, where both the front and rear axles are powered and capable of regeneration. In such architectures, the potential for energy recovery is significantly higher than in front- or rear-wheel-drive configurations, but so too is the complexity of managing the interaction between the two powertrains.
Pan and Xu’s control strategy leverages the external characteristics of the dual motors—essentially their torque-speed profiles—to determine the most efficient way to distribute braking effort. When a driver applies the brakes, the system first evaluates the vehicle’s speed, battery state of charge (SOC), and braking intensity. If the speed is above 5 km/h, the SOC is below 90%, and the braking demand is not urgent (less than 0.8 g), the system engages regenerative braking. The key innovation lies in how it then partitions the load between the front and rear motors.
The strategy is designed to prioritize energy recovery while adhering to safety constraints. It uses the estimated peak adhesion coefficients for the front and rear wheels to define upper and lower limits for the allowable braking force on each axle. If the front wheels have higher grip than the rear—a common scenario due to weight transfer during braking—the system can allocate more regenerative torque to the front axle, where it can be most effectively utilized. Conversely, if grip is more evenly distributed, the system balances the load to maintain vehicle stability.
What sets this approach apart is its ability to adapt to changing conditions in real time. For example, during a braking maneuver on a surface that transitions from wet asphalt to dry pavement, the system continuously updates its adhesion estimates and adjusts the torque distribution accordingly. This ensures that the vehicle never exceeds the available grip, even as road conditions evolve mid-deceleration. The result is a smoother, more predictable braking experience that maximizes energy recovery without requiring driver intervention.
To validate their strategy, the researchers conducted extensive simulations using CarSim and Simulink, modeling a compact four-wheel-drive EV with a curb weight of 940 kg, a wheelbase of 2.26 meters, and lithium-ion battery packs. In one test scenario, the vehicle decelerated from 40 km/h to a stop with a braking intensity of 0.3 g on a surface with a known adhesion coefficient of 0.6. Under these conditions, the system achieved a braking energy recovery rate of 65.55%—a figure that represents a significant improvement over many existing systems, which often struggle to exceed 50% under similar conditions.
The simulation also revealed the sources of energy loss, providing valuable insights for future optimization. Approximately 20% of the recoverable energy was dissipated by the mechanical brakes, primarily during the final phase of deceleration when vehicle speed dropped below 5 km/h and regenerative braking was disabled. Another 10% was lost to inefficiencies in the motor, battery, and drivetrain—losses that are inherent to any energy conversion process but could potentially be reduced through advances in component design. The remaining 5% was attributed to control system imperfections, suggesting that further refinements to the algorithm could push recovery rates even higher.
In a more complex test, the vehicle was subjected to a variable road condition, transitioning from a low-grip surface (adhesion coefficient of 0.2) to a high-grip surface (0.8) and back again, while decelerating from 80 km/h. This scenario is particularly challenging for conventional systems, which may either underutilize the available grip on the high-adhesion segment or risk instability when returning to the low-adhesion segment. The Jiangsu University strategy, however, demonstrated robust performance, with the predicted adhesion coefficients closely tracking the actual values throughout the maneuver. Although the overall energy recovery rate in this test was lower—28.90%—this was primarily due to the inherently limited grip on the low-adhesion surfaces, where mechanical braking was necessary to ensure safety. The fact that the system could accurately identify and respond to these rapid changes underscores its potential for real-world applicability.
The implications of this research extend beyond the laboratory. As EVs become more prevalent, the ability to recover energy efficiently will play an increasingly important role in reducing charging frequency, extending battery life, and lowering the total cost of ownership. For automakers, adopting a control strategy like the one proposed by Pan and Xu could provide a competitive advantage, offering drivers a more efficient and responsive braking experience. Moreover, the integration of road surface identification could enhance the performance of other vehicle systems, such as electronic stability control and adaptive cruise control, creating a more cohesive and intelligent driving environment.
From a technical standpoint, the success of this strategy hinges on the seamless integration of multiple disciplines: tire dynamics, motor control, fuzzy logic, and real-time data processing. The use of fuzzy control is particularly noteworthy, as it allows the system to handle the uncertainty and nonlinearity inherent in road-tire interactions. Unlike traditional binary logic, which operates on strict true/false conditions, fuzzy logic can manage partial truths, making it ideal for applications where inputs are imprecise or continuously changing. By mapping wheel slip and adhesion data to linguistic variables such as “small,” “medium,” and “large,” the system can make nuanced decisions that mimic human intuition.
Another key feature of the design is its emphasis on safety. The researchers implemented a “low-select” principle for the adhesion coefficient, meaning that the system always uses the lower of the two values estimated for the left and right wheels on the same axle. This conservative approach ensures that the vehicle never assumes more grip than is actually available, reducing the risk of lateral instability or spinout. Additionally, the system disables regenerative braking under high-intensity braking conditions, prioritizing immediate stopping power over energy recovery—a decision that aligns with best practices in automotive safety engineering.
The study also highlights the importance of system-level thinking in EV design. While much of the public discourse around electric vehicles focuses on battery capacity and charging speed, the reality is that efficiency gains can be found throughout the entire vehicle architecture. Regenerative braking is one of the most effective levers for improving efficiency, but its benefits are only fully realized when the control strategy is optimized for the specific characteristics of the powertrain and the driving environment. Pan and Xu’s work exemplifies this holistic approach, demonstrating how a deep understanding of both hardware and software can lead to tangible improvements in performance.
Looking ahead, the principles outlined in this research could be applied to a wide range of vehicle types, from passenger cars to commercial trucks. As sensor technology continues to advance, future iterations of the system could incorporate data from additional sources, such as cameras, radar, or even vehicle-to-infrastructure communication, to further refine adhesion estimates. Machine learning algorithms could also be used to personalize the control strategy based on individual driving habits and regional road conditions.
In an industry where incremental improvements are often celebrated as breakthroughs, the 65.55% energy recovery rate achieved by Pan and Xu’s strategy stands out as a meaningful leap forward. It is a testament to the power of academic research in driving innovation and solving real-world problems. As the global transition to electric mobility accelerates, studies like this one will play a crucial role in shaping the next generation of sustainable transportation.
The work of Pan Gongyu and Xu Shen at Jiangsu University not only advances the state of the art in regenerative braking but also serves as a model for how engineering research can directly impact the performance and efficiency of everyday technologies. Their strategy is a reminder that the future of mobility is not just about bigger batteries or faster charging, but about smarter, more adaptive systems that make the most of every joule of energy.
Published by Pan Gongyu and Xu Shen, School of Automotive and Traffic Engineering, Jiangsu University, in the Journal of Jiangsu University (Natural Science Edition), DOI: 10.3969/j.issn.1671-7775.2024.01.001