Vehicle Behavior Control for Bottleneck Lane Change Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems struggle to effectively manage vehicle behavior in bottleneck sections due to unpredictable congestion risks and lack an optimal method for lane change and acceleration control, which can lead to instability and speed decreases.
Innovation Solution
A method and apparatus using a reward function to determine vehicle behavior, incorporating internal and external rewards for target speed compliance, successful lane changes, unsafe following distance, and infeasible actions, with a decision-making model trained through reinforcement learning to optimize vehicle behavior in bottleneck sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If autonomous driving systems use conventional lane change routing methods, then lane change capability is provided, but vehicle stability and speed maintenance in bottleneck sections deteriorate
Solution Approach 1:
The patent applies parameter changes by modifying the reward function parameters to dynamically adjust vehicle behavior. The reward function includes multiple terms (speed maintenance reward, lane change reward, safety distance reward, acceleration reward) that are tuned to optimize vehicle performance in bottleneck sections, enabling the system to maintain stability while performing lane changes
Solution Approach 2:
The patent implements feedback mechanisms through the reinforcement learning framework where the vehicle's behavior is continuously evaluated based on observed outcomes. The reward function provides feedback signals that guide the decision-making model to adjust lane change timing and acceleration patterns, improving both stability and speed maintenance in bottleneck sections
2Speed
If autonomous vehicles maintain target speed in bottleneck sections, then speed compliance is improved, but congestion control and lane change optimization deteriorate
Solution Approach 1:
The patent uses parameter changes in the reward function to balance speed compliance and congestion control. The speed maintenance reward term encourages staying near target speed, while the lane change reward and acceleration reward terms promote proactive lane changes and acceleration maneuvers that improve overall congestion control efficiency
Solution Approach 2:
The patent applies dynamics by making the vehicle behavior adaptive rather than static. The reinforcement learning decision-making model dynamically adjusts lane change timing and acceleration based on real-time observations of bottleneck conditions, allowing the system to maintain speed compliance when safe while improving congestion control when opportunities arise
3Productivity
If reinforcement learning is used for decision making, then vehicle behavior optimization is improved, but system complexity increases
Solution Approach 1:
The patent manages complexity through parameter changes in the reward function design. By carefully selecting and tuning the reward terms and their weights, the system achieves effective vehicle behavior optimization without requiring overly complex model architectures, balancing performance improvement with manageable system complexity
Data Source
AI summary
Provided is a method and an apparatus for determining a vehicle behavior, and more specifically, to a method and an apparatus for determining a vehicle behavior for bottleneck congestion control in a bottleneck section. Tn apparatus for determining a vehicle behavior may include an information collection unit collecting surrounding information of a target driving vehicle from a road side unit (RSU), a vehicle observation unit obtaining observation information based on the target driving vehicle from a sensing module mounted on the target driving vehicle, a reward determination unit determining a reward for the target driving vehicle through a reward function which uses the surrounding information and the observation information, a model training unit updating and training a decision making model through the reward, and a behavior determination unit determining a behavior of the target driving vehicle by inputting the observation information into the decision making model.