Autonomous Driving Hyperparameter Control for Adaptive Trajectory Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems for autonomous vehicles face challenges in efficiently determining optimal hyperparameters for speed, obstacle avoidance, and global path following, leading to labor-intensive tuning and variable performance across different driving situations.
Innovation Solution
A processor-implemented method that generates multiple trajectories based on speed and steering, selects candidate trajectories based on obstacle presence, outputs hyperparameters using a machine learning model trained through reinforcement learning, and controls steering and speed to follow the target trajectory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual tuning of hyperparameters is used for autonomous driving, then the system can be implemented with simple structure, but the tuning process becomes labor-intensive and time-consuming
Solution Approach 1:
The system employs reinforcement learning to enable self-service automatic tuning of hyperparameters. The neural network automatically adjusts speed hyperparameter, obstacle avoidance hyperparameter, and global path following hyperparameter based on driving situations without requiring manual intervention, thus resolving the contradiction between ease of implementation and tuning time consumption.
2Device complexity
If fixed hyperparameters are used for autonomous driving, then the system structure remains simple, but performance varies across different driving situations
Solution Approach 1:
The system implements dynamic hyperparameter adjustment through reinforcement learning. The neural network continuously adapts the speed hyperparameter, obstacle avoidance hyperparameter, and global path followinghyperparameter according to real-time driving situations such as obstacle presence, road curvature, and traffic conditions, thereby achieving adaptability without excessive system complexity.
Solution Approach 2:
The system changes parameters dynamically by using reinforcement learning to adjust hyperparameter values based on driving situations. The neural network modifies the speedhyperparameter, obstacle avoidancehyperparameter, and global path followinghyperparameter according to environmental conditions, enabling the system to adapt to various driving scenarios while maintaining reasonable complexity.
3Adaptability or versatility
If reinforcement learning is used to train the machine learning model, then the system achieves adaptive performance across different scenarios, but the training process and computation become more complex
Solution Approach 1:
The system uses reinforcement learning with feedback mechanisms where the neural network receives rewards or penalties based on driving performance. The feedback loop allows the model to learn optimal hyperparameter adjustments for speed, obstacle avoidance, and global path following across different scenarios, achieving adaptability while managing training complexity through structured reward functions.
Data Source
AI summary
A processor-implemented method includes generating a plurality of trajectories for a driving situation of the moving object based on either one or both of a speed and a steering of the moving object, selecting candidate trajectories based on a presence of an obstacle among the plurality of trajectories, outputting hyperparameters related to driving of the moving object by inputting data related to the driving situation to a machine learning model, selecting a target trajectory from the candidate trajectories based on the hyperparameters, and controlling the steering and the speed such that the moving object moves along the target trajectory, wherein the hyperparameters comprise a first hyperparameter for the speed of the moving object, a second hyperparameter for a degree to which the moving object is able to avoid an obstacle, and a third hyperparameter for a global path to a destination, and wherein the hyperparameters vary while the moving object travels.