5G/6G Supercomputer-Assisted Collision Avoidance for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems lack the capability to quickly and effectively mitigate vehicle collisions due to the limitations of human reflexes and judgment at high speeds, necessitating electronic reflexes and rapid signal propagation to analyze and implement collision avoidance strategies in real-time.
Innovation Solution
A method utilizing an autonomous vehicle, access point, and supercomputer to rapidly analyze imminent collision scenarios, calculate and implement sequences of actions such as steering, braking, or acceleration to avoid or minimize collision harm, leveraging high-speed communication technologies like 5G for timely data transfer and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If human reflexes and judgment are used to avoid collisions, then the system is simple and easy to operate, but the response time is too slow at high speeds
Solution Approach 1:
The patent replaces the mechanical human reflex system with an electronic computer-based system. The onboard computer analyzes sensor data and calculates avoidance maneuvers electronically, achieving response speeds thousands of times faster than human reflexes while managing the complexity through integrated vehicle control systems.
Solution Approach 2:
The patent introduces a supercomputer as an intermediary to assist the onboard computer. When a collision is detected, the supercomputer receives data, calculates optimal avoidance strategies, and sends recommendations back to the vehicle, providing enhanced computational power without requiring the vehicle itself to have extreme processing capabilities.
2Reliability
If thousands or millions of possible actions are analyzed to find the best avoidance strategy, then the collision mitigation effectiveness is improved, but the calculation time increases
Solution Approach 1:
The patent pre-calculates and stores databases of collision scenarios, avoidance maneuvers, and outcome predictions. When a collision is detected, the system searches these pre-prepared databases rather than calculating from scratch, dramatically reducing computation time while maintaining comprehensive analysis of possible actions.
Solution Approach 2:
The patent divides the complex calculation task into segments: the onboard computer handles immediate detection and basic analysis, the access point relays data, and the supercomputer performs the heavy computational lifting for evaluating thousands of scenarios. This segmentation allows each component to operate within its capabilities without creating bottlenecks.
3Reliability
If a supercomputer is used to rapidly evaluate and select the best avoidance strategy, then the collision avoidance capability is improved, but the device complexity and cost increase
Solution Approach 1:
The patent uses an access point as an intermediary between the vehicle and supercomputer. The access point receives collision data from the vehicle, forwards it to the supercomputer, and relays the recommended avoidance strategy back to the vehicle. This distributed architecture allows the supercomputer to be located remotely, reducing the complexity and cost burden on the vehicle itself.
Solution Approach 2:
The patent designs the supercomputer system to serve multiple vehicles simultaneously through the access point network. A single supercomputer can analyze and provide avoidance strategies for multiple collision scenarios across different vehicles, amortizing the high cost and complexity of supercomputing resources across many users rather than requiring each vehicle to have its own supercomputer.
Data Source
AI summary
Traffic collisions involving autonomous vehicles can be greatly reduced by timely initiation of evasive action. However, calculating a suitable sequence of actions capable of avoiding or minimizing the collision may require the speed of a supercomputer. Therefore, disclosed is a method for the autonomous vehicle to transmit an emergency message with sensor data to a nearby access point in 5G or 6G, and the access point can forward the data to a supercomputer trained in collision avoidance. The supercomputer, millions or billions of times faster than vehicle computers, explores many sequences of actions and selects the one most likely to avoid the collision or if unavoidable, the sequence of actions that results in the least harm. Again using an exclusive channel, the supercomputer and the access point can relay the selected sequence to the autonomous vehicle, for immediate collision avoidance or harm minimization.