NexaRob News/Report and research

How does an autonomous racing car automatically adjust its speed in real time?

During Season 2 of the Abu Dhabi Autonomous Racing League (A2RL), the Mission Performance module was successfully tested on a fully autonomous EAV-25 vehicle, which automatically adjusted its speed in real time. The system does not change the parameters of the vehicle model, but directs the motion planning and control modules by adapting performance goals.

On this page

How does the automatic speed adjustment module work?

The Mission Performance module is a performance management system that does not change the parameters of the vehicle model, but adjusts performance goals for critical modules - such as motion planning and control. Instead of trying to estimate tire grip in real time, which is risky and difficult to implement, the system operates at the level of performance goals. This means that instead of modifying the physical model of the vehicle, the module tells other components: 'you can drive faster, slower or maintain the current speed - depending on the conditions.' This approach allows for greater stability and safety during dynamic,

automatic speed changes. The system uses previously defined tire warm-up routines at the beginning of the track, which prevents overheating or insufficient warming of the tires in the first seconds of the race.

The operation of the module is based on continuous monitoring of vehicle safety and dynamics metrics - sector by sector. Each part of the track is analyzed individually for driving conditions, allowing the system to decide in real time whether to reduce, maintain or increase the level of performance. In this way, the module gradually approaches the maximum permissible speed value, avoiding excessive risks and ensuring a smooth race.

Yas Marina track tests - a real verification of effectiveness.

The effectiveness of the solution was confirmed during Season 2 of the Abu Dhabi Autonomous Racing League (A2RL) on the Yas Marina track. This is where the EAV-25 vehicle operated - a fully autonomous Dallara Superformula - which used the Mission Performance module in real conditions. The study concerned not only simulations, but also actual driving on the race track with full dynamic loads and changing track conditions.

The results indicate that the system can effectively manage the race - from the initial tire warm-up to the gradual increase in driving speed. This allows the car to achieve faster lap times without compromising safety, which is a key objective in autonomous racing projects. However, tests have not demonstrated the possibility of fully automating all aspects of driving - the module functions as an aid for other systems rather than replacing them.

Autonomous racing car EAV-25 on the Yas Marina track.
Tests on the Yas Marina track - real verification of effectiveness - illustrative visualization.

Why is this important for the future of autonomous mobility?

The Mission Performance solution shows how high performance can be achieved in autonomous systems without the need to continuously model complex physical factors - such as grip force. Instead of trying to 'understand' every moment of driving, the system operates at the level of goals and adaptation. This approach can be applied to other applications where safety and smoothness are equally important - for example, in autonomous logistics or urban transport.

However, it is worth emphasizing that the results apply only to one type of vehicle (EAV-25) and one track (Yas Marina). There is no information on the scalability of the solution to other classes of vehicles, weather conditions, or different tracks. Therefore, it is not yet possible to talk about a general solution for the entire industry - this is a specific implementation in a racing context.

Limitations and what remains unclear.

Although the module has been tested on a real track, its effectiveness has not been compared with other approaches - for example, systems based on real-time grip estimation. There is no data on how the system performs in rainy conditions, changes in track temperature, or large differences between sectors. Furthermore, it is unknown whether the module can operate without access to previously defined tire heating profiles - whether it is flexible in non-standard starting conditions.

It is also worth noting that the document does not contain data on the system's reaction time or its computational performance. There is a lack of information on how often the module updates its decisions and whether it can operate under limited computing power - which is important for commercial applications.

It is worth emphasizing that, although the Mission Performance module shows potential in a racing context, its application outside this environment requires further research. The lack of data on computational performance and system reaction time limits its practical application in real-world conditions, where decisions must be made in milliseconds. In addition, the lack of comparison with other approaches - such as real-time grip estimation - makes it impossible to assess its unique value compared to other solutions. The system works best in controlled conditions and with predefined profiles, which means limited flexibility in dynamic or non-standard scenarios.

Therefore, its role remains supportive - it does not replace full autonomy, but enables smooth and safe adaptation of the driving pace to current track conditions, which is a key step towards more advanced autonomous systems. In the future, this approach may be extended to other areas of mobility, but only after further validation and adaptation to diverse operating conditions.

Editorial transparency

Sources and reference materials

The article was prepared by NexaRob based on an analysis of available source materials. The following materials were used to verify information and expand the context.

1source material
1primary
1publicly shown
  1. Primary sourceResearchData

    Mission Performance: Automatic and Adaptive Race Pace Progression for Autonomous Racing

    arXiv Robotics)cs.RO)arxiv.org

How to read this section? Sources are materials used during research and verification. The article is an original NexaRob report, not a reprint of the indicated publications.

Share the material
Go to sources
EnglishEN