Self-Driving Car Trajectory Planning for Dynamic Object Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving cars face challenges in navigating safely around dynamic objects, particularly when these objects are in neighboring lanes, as existing systems struggle to effectively predict and respond to their trajectories, increasing the risk of collisions.
Innovation Solution
An electronic device with modules for data processing, including perception, trajectory prediction, dynamic distance calculation, and movement planning, is used to generate control data for self-driving cars. This device analyzes sensor data to determine the trajectories of dynamic objects and calculates a dynamic index to identify the most risk-averse trajectory, reducing collision risk by adjusting the car's path in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the SDC uses sensor systems to collect and analyze motion data of dynamic objects to determine estimated trajectories, then the risk of collision with dynamic objects is reduced, but the device complexity and computational requirements increase
Solution Approach 1:
The system performs preliminary trajectory estimation for dynamic objects by analyzing sensor data and predicting future positions. This advance preparation allows the path planning system to proactively avoid potential collision zones rather than reacting to actual collisions, thereby reducing collision risk while managing computational complexity through early intervention
Solution Approach 2:
The patent introduces an intermediate trajectory estimation layer between sensor data collection and path planning. This intermediary component processes raw sensor data to generate predicted trajectories of dynamic objects, which then inform the path planning algorithm. This mediator simplifies the overall system architecture by preprocessing information before it reaches the decision-making layer
2Reliability
If the system calculates dynamic indexes and adjusts the SDC trajectory in real-time based on predicted object trajectories, then collision risk is reduced, but the computational processing time and energy consumption increase
Solution Approach 1:
The system performs partial trajectory calculations by focusing computational resources on dynamically relevant parameters and time horizons. Rather than calculating complete trajectories for all possible future scenarios, the system computes only the necessary portion of trajectory data needed for immediate path planning decisions, reducing overall computational energy consumption while maintaining safety
Solution Approach 2:
The system updates trajectory predictions and recalculates dynamic indexes at periodic intervals rather than continuously. This periodic recalculation approach balances safety requirements with energy consumption by performing intensive computational tasks at scheduled moments while using interpolated or cached data between updates, thereby reducing overall energy usage
3Reliability
If the system generates and evaluates multiple candidate trajectories to identify the most risk-averse path, then collision avoidance capability is improved, but the productivity and response time of the path planning process deteriorate
Solution Approach 1:
The system dynamically adjusts the number of candidate trajectories generated and the depth of evaluation based on real-time conditions. In high-risk environments with multiple dynamic objects, the system generates more candidate trajectories to ensure safe path selection. In low-risk situations, it reduces the number of candidates to improve processing speed. This dynamic adaptation resolves the contradiction between thoroughness and speed
Solution Approach 2:
The system maintains continuous path planning operations by implementing a pipeline that generates, evaluates, and selects trajectories without interruption. Multiple candidate trajectories are generated and evaluated in a continuous flow, allowing the system to maintain high productivity while still performing comprehensive collision avoidance analysis through parallel processing and optimized evaluation sequences
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and devices for generating data for controlling a Self-Driving Car (SDC) are disclosed. The method includes: (i) acquiring a predicted object trajectory for an object, (ii) acquiring a set of anchor points along the lane for the SDC, (iii) for each one of the set of anchor points, determining a series of future moments in time when the SDC is potentially located at the respective one of the set of anchor points, thereby generating a matrix structure including future position-time pairs, (iv) for each future position-time pair in the matrix structure, using the predicted object trajectory for determining a distance between a closest object to the SDC as if the SDC is located at the respective future position-time pair, and (v) storing the distance between the closest object to the SDC in association with the respective future position-time pair in the matrix structure.