Vehicle Network Attack Detection via Multi-Vehicle LIDAR Voting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inaccurate or incomplete LIDAR data from autonomous or semi-autonomous vehicles can impair safe operation, particularly due to inconsistencies in object detection and classification, which may be exacerbated by cyber-attacks manipulating sensor data or occlusions, affecting vehicle navigation and collision avoidance systems.
Innovation Solution
A computer system that receives object data from multiple vehicles and infrastructure sensors, identifies inconsistencies using a voting technique based on majority object data, and performs actions such as ignoring suspicious data, actuating vehicle brakes, or outputting messages to ensure safe operation by determining occlusions and misclassifications, thereby maintaining vehicle safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LIDAR data from a single vehicle sensor is used for navigation, then the system is simpler and faster, but the reliability and accuracy of object detection deteriorate due to potential inaccuracies, occlusions, or cyber-attacks
Solution Approach 1:
A central server acts as an intermediary to receive LIDAR data from multiple vehicles, perform consistency checks using voting techniques, and distribute verified object information back to vehicles. This mediator handles the complex verification process centrally, allowing individual vehicles to maintain simpler local systems while achieving high reliability through collective validation.
Solution Approach 2:
The system merges LIDAR data from multiple vehicle sensors and infrastructure sensors into a unified dataset. By combining data sources and applying voting techniques, the system achieves more reliable object detection than any single sensor could provide alone, while the central server manages the complexity of integrating and verifying these multiple data streams.
2Measurement precision
If LIDAR data from multiple vehicles is collected and verified using voting techniques, then the accuracy and reliability of object detection improves, but the data processing time and computational complexity increases
Solution Approach 1:
The system continuously collects and pre-processes LIDAR data from multiple vehicles and infrastructure sensors in real-time, maintaining an up-to-date database of object information before conflicts or inconsistencies arise. This preliminary data gathering and continuous verification process reduces the time needed for critical decision-making, as the foundation for voting and consistency checks is already established.
Solution Approach 2:
The central server implements continuous feedback loops where LIDAR data from multiple sources is constantly compared and verified. When inconsistencies are detected, the system quickly identifies and corrects them through voting mechanisms, providing real-time feedback to vehicles about verified object locations. This continuous feedback process maintains high precision without significant time loss by operating continuously rather than reactively.
3Object-affected harmful factors
If inconsistent object data from one vehicle is used, then the system operates with fewer sensors, but the safety and navigation accuracy deteriorates due to potential cyber-attacks or sensor failures
Solution Approach 1:
The central server serves as a trusted intermediary that receives LIDAR data from multiple vehicles, performs consistency checks to detect potential cyber-attacks or sensor failures, and distributes verified information back to vehicles. This intermediary layer protects the system from harmful data without requiring each individual vehicle to implement complex attack detection algorithms.
Solution Approach 2:
The system implements continuous feedback where the central server monitors LIDAR data from multiple vehicles for inconsistencies that may indicate cyber-attacks. When anomalies are detected, the feedback mechanism identifies and isolates the problematic data source, allowing the system to maintain safety by relying on consistent data from other vehicles while managing the complexity of multi-vehicle coordination centrally.
Data Source
AI summary
A computer includes a processor and a memory. The memory stores instructions executable by the processor to receive respective object data including an object location from each of three or more vehicles, based on determining an inconsistency of the object data received from a first vehicle with respect to respective object data of the other vehicles, perform an action.