Ultrasonic Object Detection Grouping for Complete Surround Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for generating environmental representations using ultrasonic sensors in autonomous machines require excessive computing resources and latency, often failing to detect objects that are not closest or within static sensor groups, leading to incomplete navigation information.
Innovation Solution
Group object detections based on sensor configurations, such as order and distance thresholds, to dynamically generate groups of ultrasonic data, allowing for efficient processing and detection of objects at various distances from the machine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a brute force technique is used to process every combination of object detections, then all object locations can be determined, but a large amount of computing resources and latency are required
Solution Approach 1:
The patent segments the set of all object detections into multiple groups based on spatial relationships and distance thresholds. Instead of processing every possible combination of detections, the system divides them into manageable groups where only detections within the same group are processed together. This segmentation dramatically reduces the computational complexity from exponential to linear or polynomial time while still identifying all object locations through hierarchical group processing.
2Productivity
If a closest object technique is used to process only the closest object detections, then computing resources are reduced, but objects located behind other objects cannot be detected
Solution Approach 1:
The patent performs preliminary grouping of object detections based on spatial criteria and distance thresholds before processing. By pre-organizing detections into groups where multiple detections in the same group can indicate different objects (including those behind others), the system ensures that behind-object detections are not discarded. This preliminary action preserves detection completeness while enabling efficient group-based processing that reduces computational resources.
3Productivity
If a dividing technique is used to group sensors based on their locations, then computing resources are reduced and objects at various distances can be detected, but objects located between sensor groups cannot be detected
Solution Approach 1:
The patent implements dynamic group formation where object detection groups are not fixed to specific sensor locations but are instead formed based on the spatial relationships among detected objects themselves. The grouping criteria include distance thresholds and relative positions that allow detections from different sensor groups to be combined when they indicate objects in between sensor locations. This dynamic approach ensures objects between sensor groups are detected while maintaining computational efficiency.
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
Reduces computing resources needed while detecting all surrounding objects, including those behind or between sensor groups, providing comprehensive navigation information.
Implementation Method 1
ultrasonic data generated using ultrasonic sensors of a machine
Implementation Method 2
Based on the one-dimensional outputs represented by ultrasonic data
Data Source
AI summary
In various examples, associating object detections for sensor data processing for autonomous and semi-autonomous systems and applications is described herein. Systems and methods described herein may group object detections (e.g., echoes, etc.) that are detected using multiple sensors (e.g., ultrasonic sensors, sonar sensors, etc.) and then use the groupings to process the object detections to perform one or more tasks, such as object or feature detection. In some examples, the object detections are grouped using one or more configurations, such as an order associated with analyzing sensor data generated using the sensors and/or threshold distances associated with determining that object detections are associated with the same object. Additionally, a respective group may be generated for one or more (e.g., each) detected object such that the locations of the object surrounding the machine may be determined.