Multi-Sensor Target Detection With Time-Series Reliability Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing target detection systems struggle to accurately determine the reliability of targets when they move out of the sensing range of multiple sensors, leading to potential inaccuracies in automatic braking decisions.
Innovation Solution
A target detection device that integrates detection information from multiple sensors, including cameras and radars, to manage integration history in time series and determine reliability based on this history and detection information, using a reliability determination unit to classify reliability into 'confirmed', 'past', and 'split' categories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reliability determination is based only on simultaneous detection by multiple sensors, then the determination process is simple, but reliability cannot be appropriately determined when targets move out of sensor ranges
Solution Approach 1:
The system performs preliminary detection and classification using the camera before radar detection is needed. The camera pre-identifies potential targets and their types, which are then tracked by the radar even when the target moves out of the camera's field of view. This preliminary action by the camera enables reliable target type determination to be maintained throughout the target's movement trajectory.
Solution Approach 2:
The camera acts as an intermediary that provides initial target type information. When the radar detects a target, it uses the camera's previous classification results as reference to determine the target type, even when simultaneous camera-radar detection is not occurring. This intermediary role of the camera enables reliable target type determination beyond the limitations of simultaneous detection.
2Reliability
If only simultaneous detection by multiple sensors is used, then the processing is straightforward, but target tracking continuity is lost when targets move out of sensing ranges
Solution Approach 1:
The camera performs preliminary target identification and type classification, creating a record that persists even when the target moves out of the camera's immediate sensing range. This preliminary classification enables the system to maintain target type information over time, preventing loss of detection coverage when targets transition between sensor ranges.
Solution Approach 2:
The system transitions from relying solely on spatial overlap of sensor fields of view to utilizing temporal dimension for target tracking. By integrating camera classification results over time with radar detection data, the system maintains continuous target type determination even when targets move in and out of simultaneous detection zones, effectively adding the time dimension to detection coverage.
3Measurement precision
If automatic braking decisions are made without reliable target classification, then response time is fast, but braking accuracy deteriorates
Solution Approach 1:
The camera performs preliminary target type classification before automatic braking decisions are required. This pre-classification of targets into categories (pedestrian, vehicle, animal, etc.) is stored and readily available when braking decisions must be made, enabling accurate target identification without adding decision-making time.
Solution Approach 2:
The system uses feedback from camera classification results to inform and validate radar-based target type determination. When the radar detects a target, it references the camera's classification feedback to confirm or refine the target type, enabling accurate target classification that supports precise automatic braking decisions within required response times.
Data Source
AI summary
Reliability of a target is appropriately determined. A target detection device detects a target around a vehicle based on sensor information measured by a plurality of sensors. The device includes: a first sensor; a second sensor; an integration processing unit that integrates detection information of the target of the first sensor and detection information of the target of the second sensor for each cycle; an integration history management unit that manages an integration result of the integration processing unit as an integration history in time series; and a reliability determination unit that determines reliability of a type of the target based on the integration history and the pieces of detection information of the target of the first sensor and the second sensor. The detection information of the target of the first sensor includes type information of the target.