Vehicle Object Tracking With Single-Sensor Direct Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking systems face issues with reaction time optimization due to latency, bandwidth limitations, and unnecessary data processing efforts, particularly when combining sensor data from multiple sensors, leading to delayed responses and increased energy consumption.
Innovation Solution
A method for sensor-based object tracking that distinguishes between single sensor objects and multisensor objects, using sensor-specific identifiers and direct data updating without additional filtering, and employs a filter-based approach only for multisensor objects, thereby reducing redundant processing and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data from multiple sensors is combined using traditional fusion methods, then measurement precision and reliability are improved, but processing time and latency increase
Solution Approach 1:
The patent segments object tracking into two distinct categories: single-sensor objects and multi-sensor objects. This segmentation allows different processing strategies to be applied to each category, avoiding the need to process all objects through the same complex multi-sensor fusion pipeline, thereby reducing overall processing latency while maintaining precision where needed.
Solution Approach 2:
The patent applies local quality by treating single-sensor objects and multi-sensor objects differently in terms of processing depth and methods. Single-sensor objects receive streamlined processing, while multi-sensor objects receive enhanced fusion processing, optimizing the balance between processing speed and measurement precision for each local case.
2Reliability
If multiple sensors are used for comprehensive object detection, then reliability is improved, but energy consumption increases
Solution Approach 1:
The patent applies partial action by selectively engaging multiple sensors only when necessary (for multi-sensor objects) rather than continuously processing all sensor data through full fusion algorithms. This reduces overall energy consumption while maintaining reliability for objects that require multi-sensor verification.
Solution Approach 2:
By segmenting objects into single-sensor and multi-sensor categories, the system can allocate computational resources more efficiently, processing single-sensor objects with lower energy consumption methods and reserving intensive multi-sensor fusion for only those objects that require enhanced reliability.
3Measurement precision
If traditional filter-based processing is applied to all object data, then measurement precision is improved, but processing complexity and time increase
Solution Approach 1:
The patent segments the object tracking process into two distinct pathways: one for single-sensor objects that uses simpler processing, and one for multi-sensor objects that uses filter-based fusion processing. This segmentation reduces overall processing complexity by avoiding the application of complex filters to all data uniformly.
Solution Approach 2:
The patent applies local quality by applying filter-based processing only to multi-sensor objects where it is most needed, while using simpler processing methods for single-sensor objects, thereby optimizing the balance between measurement precision and processing complexity for each case.
Data Source
AI summary
A method and an object tracking unit for sensor-based object tracking and a correspondingly configured motor vehicle. In the method, objects which are only detected by a single one of multiple sensors are identified as single sensor objects. Corresponding object tracks, in which object data describing the respective object are stored, are then updated directly based on sensor data without using a predetermined filter. Other object tracks, to which no single sensor object was assigned, are updated based on the sensor data using the predetermined filter.