Vehicle Sensor Fusion Track Switching for Sensor Failure Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face control malfunctions when any one of their sensor fusion devices, such as cameras, radars, or Lidars, fails to recognize objects, leading to a lack of redundancy in object detection.
Innovation Solution
A vehicle system that processes image data, radar data, and Lidar data to generate a first sensor fusion track, calculates the reliability of each sensor, and switches to a second sensor fusion track based on a predetermined threshold, allowing for redundancy by using alternative sensor data for braking and deceleration control, even if one sensor fails.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor fusion is used for object detection, then measurement precision is improved, but reliability deteriorates when a sensor fails
Solution Approach 1:
The system pre-establishes multiple sensor fusion tracks (first track using all sensors, second track using subset of sensors) and calculates reliability thresholds in advance. When a sensor failure is detected, the system can immediately switch to the pre-prepared alternative track without interruption, cushioning against the reliability deterioration that would otherwise occur during sensor failure.
Solution Approach 2:
The system dynamically changes the composition parameters of the sensor fusion track based on sensor reliability. When a sensor fails, the system transitions from the first sensor fusion track (using all sensors for high precision) to the second sensor fusion track (using remaining sensors with calculated reliability), adapting the system configuration to maintain operation under changed conditions.
2Reliability
If alternative sensor fusion track is prepared for redundancy, then reliability is improved during sensor failure, but device complexity increases
Solution Approach 1:
The sensor fusion system is segmented into multiple independent tracks: the first sensor fusion track uses all sensors for normal operation, while the second sensor fusion track uses a subset of sensors for failure scenarios. This segmentation allows the system to maintain reliability through redundancy without requiring complete reconfiguration, as each track is a self-contained processing path.
Solution Approach 2:
The system dynamically selects between the first and second sensor fusion tracks based on sensor operational status. The controller adjusts which track is active in real-time, enabling the system to maintain high reliability during sensor failures while avoiding the permanent complexity of maintaining all possible sensor combinations simultaneously.
3Reliability
If braking amount is limited when using alternative sensor fusion track, then safety is maintained during sensor failure, but productivity decreases
Solution Approach 1:
The controller dynamically adjusts the braking amount parameter based on the active sensor fusion track. When the second track is active (indicating sensor failure), the braking amount is limited to a predetermined ratio to maintain safety margins. When the first track is active (all sensors functioning), full braking capability is available, maximizing productivity and response effectiveness.
Data Source
AI summary
A vehicle includes a camera provided to obtain image data, radar provided to obtain radar data, Lidar provided to obtain Lidar data, and a controller configured to process the image data, the radar data, and the Lidar data to generate a first sensor fusion track, where the controller calculates reliability of at least one sensor in which an event does not occur among a plurality of sensors including the camera, the radar and the Lidar when the event for the sensor is detected, and changes from the first sensor fusion track to a second sensor fusion track based on the at least one sensor when the reliability is greater than or equal to a predetermined threshold, and where the controller is configured to control a braking amount or a deceleration amount of the vehicle based on the first sensor fusion track or the second sensor fusion track.