UWB-Camera Object Detection Fusion for Accurate Vehicle Depth Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object detection systems for autonomous vehicles face challenges with complexity, computational intensity, and power consumption, particularly when using LiDAR and radar sensors. Additionally, ultra-wide band (UWB) sensor networks have limited detection range and require the target object to be equipped with a UWB sensor. Non-stereo camera systems also struggle with accurate depth estimation.
Innovation Solution
An object detection system that combines an ultra-wide band (UWB) sensor network with a non-stereo camera system. The system includes three or more anchors mounted on the vehicle and a tag on the target object, providing real-time distance measurements. The non-stereo camera captures image data, and controllers estimate camera-based and UWB-based locations of the target object using Bayesian filtering, specifically a Kalman filter, to fuse the data and improve location estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LiDAR sensors are used for object detection and distance ranging, then measurement precision is improved, but device complexity and computational requirements increase
Solution Approach 1:
The patent combines UWB sensor network for distance measurement with non-stereo camera system for object detection, merging two different sensing modalities to achieve accurate depth estimation without requiring complex LiDAR systems. The UWB anchors and tags work together with the camera to provide complementary information for location estimation.
Solution Approach 2:
The patent introduces a calibration procedure that uses UWB distance measurements as an intermediary to calibrate the camera's depth estimation. The UWB system acts as a mediator to provide ground truth distance information that corrects the camera's inherent depth inaccuracies, enabling the camera to achieve LiDAR-level precision without the computational burden.
2Length of stationary object
If radar sensors are used for object detection, then detection range is improved, but power consumption increases
Solution Approach 1:
The patent segments the sensing function across multiple low-power components: UWB anchors mounted on the vehicle, UWB tags on target objects, and a non-stereo camera. Each component operates at low power consumption, yet together they achieve detection capabilities that extend beyond what a single low-power sensor could provide, avoiding the high power consumption of radar systems.
3Device complexity
If UWB sensor network is used for object detection, then device complexity is reduced, but detection range and measurement precision are limited
Solution Approach 1:
The patent makes the UWB system multi-functional by using it for both distance measurement (providing depth information) and object detection (identifying target locations). The same UWB anchor-tag infrastructure serves dual purposes: calibrating camera depth and detecting object positions, thereby maintaining simplicity while improving precision through the complementary camera system.
4Device complexity
If non-stereo camera system is used for object detection, then device complexity is reduced, but depth estimation accuracy deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where UWB distance measurements are used to calibrate and correct the camera's depth estimation. The calibration procedure establishes a relationship between camera image data and actual distances, and this calibrated model continuously corrects the camera's depth estimates during operation, enabling accurate depth measurement without stereo vision.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces complexity and computational requirements compared to LiDAR-based systems and decreases power consumption compared to radar-based systems. It also provides improved depth estimation and location accuracy for target objects, even with non-stereo camera systems.
Implementation Method 1
each anchor sends and receives sensor signals that indicate real-time distances between each anchor and the tag
Data Source
AI summary
An object detection system for a vehicle that estimates a location of a target object located in an environment surrounding the vehicle includes an ultra-wide band (UWB) sensor network including three or more anchors mounted to the vehicle in wireless communication with a tag mounted to the target object, a non-stereo camera system that captures image data representing the target object located in the environment surrounding the vehicle, and one or more controllers in electronic communication with the UWB sensor network and the non-stereo camera system. The one or more controllers includes one or more processors that execute instructions to fuse together a camera-based location of the target object and a UWB-based location of the target object by a Bayesian filter to estimate the location of the target object.