Adaptive Race Driving Control for Dynamic Path Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating race tracks and similar route-based scenarios due to the need to dynamically adjust the path based on speed changes and optimize driving for optimal lap times, while maintaining control and efficiency.
Innovation Solution
An adaptive autonomous driving algorithm that uses a processor to receive environmental data, determine a ground truth path, compare current vehicle location to the path, and adjust steering, throttle, and brake values to maintain the path and optimize speed, fuel efficiency, and tire wear.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the vehicle drives faster to improve lap time, then productivity is improved, but the optimal path changes continuously making navigation more difficult
Solution Approach 1:
The system dynamically adjusts the optimal path based on current vehicle speed. As speed changes, the navigation algorithm recalculates and updates the ground truth path to reflect the new optimal route, allowing the vehicle to maintain optimal navigation at varying speeds rather than following a static path
Solution Approach 2:
The system continuously monitors vehicle speed and uses this feedback to update the optimal path calculation. The navigation algorithm receives speed information and adjusts the ground truth path accordingly, creating a closed-loop system where speed changes trigger path optimization updates
2Measurement precision
If deep learning and CNNs are used to understand the environment, then measurement precision is improved, but the system cannot handle situations where expected movement does not match actual movement
Solution Approach 1:
The system implements continuous feedback by comparing extrapolated vehicle location (based on operating parameters) with actual GPS location. When deviations are detected, the system generates corrective steering commands to realign the vehicle with the ground truth path, ensuring reliable navigation even when environmental understanding fails to predict actual movement
Solution Approach 2:
The system introduces an intermediary verification layer that compares predicted position with actual position. This intermediary check acts as a safety mechanism that detects discrepancies between expected and actual vehicle behavior, triggering corrective actions independent of the deep learning environmental understanding
Data Source
AI summary
The present disclosure provides systems and methods for determining autonomous vehicle navigation settings and/or adjustments. In some aspects, vehicles may comprise an environmental sensor, processor, a navigation controller, and software causing the systems to utilize current location information, extrapolated location information, and a priori path locations, along with vehicle control settings, to output updated steering, braking, and throttling settings. In some aspects, methods may be utilized that reliably determine deviation from a known path that would be caused by current vehicle settings, and use the deviation to adjust the vehicle settings to improve following of the path, while optimizing vehicle attributes like speed, fuel economy, tire wear, or the like as able given primary navigation goals.