Driving Smarter: How Acceleration Variance Shapes EV Efficiency
As the global automotive industry accelerates toward electrification, a growing body of research is shifting focus from vehicle hardware to human behavior—specifically, how drivers influence the energy consumption of electric vehicles (EVs). While advancements in battery technology and powertrain efficiency continue to dominate headlines, a recent study published in the Chinese Journal of Automotive Engineering reveals that one of the most significant factors in EV energy use may not be found under the hood, but rather in the way the vehicle is driven.
The study, led by Wang Yingdi from GAC Automotive Research & Development Center and co-authored by Li Qingfeng, Wang Shi, Liu Wei, Wang Boyuan, and Xiao Jianhua from Tsinghua University, offers a comprehensive analysis of real-world driving patterns and their impact on battery energy consumption. Based on data collected from a popular electric SUV model, the researchers identify acceleration variance as the most critical factor affecting energy efficiency—more so than average speed, driving duration, or even ambient conditions.
This finding challenges the common perception that higher speeds are the primary culprit behind reduced EV range. Instead, the research suggests that the smoothness and consistency of driving behavior play a far more decisive role. Drivers who frequently accelerate and brake aggressively, even at moderate speeds, are significantly increasing their vehicle’s energy demands. In contrast, those who maintain steady speeds and anticipate traffic flow can extend their driving range without sacrificing much time.
The implications of this research are profound. As governments around the world push for greater adoption of EVs to meet climate targets, consumer confidence in vehicle range remains a major barrier. Many drivers still experience “range anxiety,” particularly on long trips or in areas with sparse charging infrastructure. Automakers have responded by increasing battery capacity and improving thermal management systems, but these solutions come with higher costs and environmental trade-offs.
This new study suggests that a more sustainable and cost-effective path to improved efficiency lies in driver education and behavior modification. By understanding the physics behind energy loss during acceleration and deceleration cycles, drivers can adopt strategies that align with the unique characteristics of electric propulsion systems.
The research team began by analyzing a dataset of real-world driving trips conducted during the summer months in urban, suburban, and highway environments. They extracted 13 key driving parameters, including average speed, speed standard deviation, acceleration standard deviation, time spent in different speed ranges, and air conditioning usage. To isolate the impact of driving behavior, the study excluded energy consumed by HVAC and other auxiliary systems, focusing solely on the battery-to-wheel energy used for propulsion.
Correlation analysis revealed several expected trends: higher average speeds, greater speed fluctuations, and increased time spent at high velocities were all positively correlated with higher energy consumption. Similarly, aggressive driving—defined as acceleration or deceleration exceeding 1.5 m/s²—was linked to elevated energy use. However, the most striking finding emerged from the dimensionality reduction analysis using Sliced Inverse Regression (SIR), a statistical method that identifies the most influential variables in complex datasets.
Among all the parameters, acceleration standard deviation (δa) emerged as the single most influential factor. This metric, which quantifies the variability of acceleration over time, proved to be a stronger predictor of energy consumption than average speed or any other speed-related statistic. In essence, how consistently a driver maintains speed matters more than how fast they travel.
The researchers explain this phenomenon through the dual nature of EV energy dynamics. During acceleration, electrical energy from the battery is converted into kinetic energy, with losses occurring in the motor, inverter, and battery itself. During deceleration, much of that kinetic energy can be recovered through regenerative braking and returned to the battery. However, this energy recovery is not 100% efficient—typically ranging from 60% to 80% depending on the system and driving conditions.
Every acceleration-deceleration cycle therefore results in a net energy loss, proportional to the magnitude of the change in speed. The more frequently and sharply a driver accelerates and brakes, the more energy is lost in these conversion cycles. This is particularly pronounced at lower speeds, where aerodynamic drag is minimal and the primary energy demand comes from overcoming inertia.
To further explore these dynamics, the team developed a one-dimensional vehicle simulation model in MATLAB/Simulink. The model included detailed representations of the vehicle control unit, motor controller, battery management system, and drivetrain, allowing for accurate prediction of energy flow under various driving conditions. The model was validated against real-world test data, showing average errors below 5% in most scenarios.
Using this simulation platform, the researchers examined three typical driving situations: steady-state cruising, frequent acceleration and deceleration (as in city traffic), and hill climbing and descent.
In the first scenario, steady-speed driving, the results confirmed the well-known relationship between speed and energy consumption. At very low speeds (around 30 km/h), energy use is relatively high due to low motor efficiency at light loads. As speed increases, motor efficiency improves, leading to a minimum in energy consumption. Beyond this point, aerodynamic drag—which increases with the square of speed—becomes dominant, causing energy use to rise sharply.
For the tested SUV, the most energy-efficient speed was approximately 30 km/h. However, the researchers emphasize that this does not mean drivers should aim to travel at this speed. Instead, they introduce the concept of a “time-energy trade-off,” recognizing that drivers value both efficiency and travel time.
When plotted on a logarithmic scale, the relationship between speed, energy use, and travel time reveals an optimal range between 70 and 80 km/h. Within this range, the time saved by driving faster outweighs the additional energy consumed. Beyond 80 km/h, the energy penalty grows faster than the time benefit, making higher speeds inefficient from a holistic perspective.
This insight has practical implications for highway driving. Many EV drivers, especially those with low battery levels, instinctively reduce their speed to maximize range. While this strategy works, the study suggests that reducing speed from 120 km/h to 100 km/h yields diminishing returns. A more effective approach is to maintain a steady pace around 80 km/h, avoiding unnecessary acceleration and using cruise control when possible.
The second scenario—frequent acceleration and deceleration—mirrors urban driving conditions, where stop-and-go traffic is common. Here, the simulation results were dramatic. At an average speed of 20 km/h, increasing the acceleration intensity from mild (0.28 m/s²) to aggressive (1.39 m/s²) caused energy consumption to rise by 237% compared to steady driving at the same average speed.
Even at higher speeds, the impact was significant. At 60 km/h, aggressive acceleration increased energy use by 91%. However, the researchers note an important nuance: at higher speeds, the relative impact of acceleration variance decreases. This is because aerodynamic drag becomes the dominant energy consumer, and the motor operates in a more efficient range. As a result, brief bursts of acceleration—such as when merging or overtaking—have a smaller effect on overall efficiency than similar maneuvers in city traffic.
This finding supports a nuanced approach to eco-driving. In congested urban areas, the priority should be smoothness: maintaining a safe following distance, anticipating traffic signals, and using gentle pedal inputs. On highways, where traffic flow is more consistent, occasional acceleration is less detrimental, and drivers can focus more on maintaining an optimal cruising speed.
The third scenario—hill driving—adds another layer of complexity. Conventional wisdom might suggest that climbing hills consumes more energy, while descending saves it. However, the study shows that driver behavior can significantly alter this balance.
When ascending a hill, reducing throttle input allows the vehicle to slow down gradually, converting some of the kinetic energy into potential energy. When descending, maintaining a light throttle or coasting allows gravity to assist propulsion, minimizing the need for regenerative braking. Because regenerative braking is not 100% efficient, minimizing its use—especially on long descents—can improve overall efficiency.
The simulations showed that for a 4% grade, a strategy of reducing speed on the ascent and allowing it to increase on the descent could reduce energy consumption by up to 17% compared to maintaining a constant speed. However, this benefit comes with a trade-off: longer travel time. At lower speeds (e.g., 40 km/h), the time penalty outweighs the energy savings, making constant-speed driving the better choice. At higher speeds (e.g., 80 km/h), the energy savings are substantial enough to justify the extra time.
This leads to a key insight: the optimal driving strategy depends on context. There is no one-size-fits-all approach to eco-driving. Instead, drivers must adapt their behavior to the specific conditions they face.
The study also highlights the importance of vehicle design in enabling efficient driving. Features such as single-pedal driving, predictive cruise control, and navigation-integrated energy management systems can help drivers make better decisions. For example, a navigation system that knows the route profile can suggest optimal speeds for upcoming hills or recommend charging stops based on real-time energy consumption.
Moreover, the findings have implications for vehicle testing and certification. Current driving cycles, such as the NEDC and CLTC, do not fully capture the variability of real-world driving. Incorporating metrics like acceleration variance into official test procedures could provide a more accurate picture of real-world efficiency and help consumers make better-informed choices.
From a policy perspective, the research underscores the value of driver education programs. Many countries already offer eco-driving courses for commercial fleets, but these are less common for private consumers. Integrating eco-driving principles into driver’s license training or offering incentives for completing efficiency courses could yield significant energy savings at scale.
Automakers, too, have a role to play. While performance-oriented driving modes may appeal to some buyers, the study shows that “sport” modes, which encourage rapid acceleration, can increase energy consumption by over 10% compared to “eco” modes. Making eco-driving the default setting, or providing real-time feedback on energy use, could nudge drivers toward more efficient behavior.
The study also touches on the psychological aspect of driving. Surveys cited in the paper indicate that EV drivers are more conscious of energy consumption than their internal combustion engine counterparts. Many are willing to accept longer travel times in exchange for greater efficiency, especially when battery levels are low. This mindset shift—from maximizing speed to optimizing efficiency—is a hallmark of the EV era.
Looking ahead, the integration of connected and autonomous technologies could further enhance eco-driving. Vehicles that communicate with each other and with infrastructure could coordinate speeds to minimize stops and starts. Autonomous driving systems could optimize acceleration and braking profiles in real time, achieving levels of efficiency beyond what most human drivers can achieve.
However, the researchers caution against over-reliance on automation. Even with advanced systems, human drivers will remain responsible for many trips for the foreseeable future. Therefore, equipping them with the knowledge and tools to drive efficiently is essential.
In conclusion, the study by Wang Yingdi and colleagues offers a data-driven framework for understanding and improving EV efficiency. It moves beyond simplistic advice like “drive slower” to provide nuanced, context-specific guidance grounded in physics and real-world data. By focusing on acceleration variance as the key lever, it empowers drivers to make informed choices that balance energy use, travel time, and comfort.
As the world transitions to electric mobility, such research will play a crucial role in maximizing the environmental and economic benefits of the technology. The message is clear: the most powerful tool for extending EV range may not be a bigger battery, but a smarter driver.
Wang Yingdi, Li Qingfeng, Wang Shi, Liu Wei, Wang Boyuan, Xiao Jianhua, GAC Automotive Research & Development Center, Tsinghua University, Chinese Journal of Automotive Engineering, DOI: 10.3969/j.issn.2095-1469.2024.03.19