Ultrasonic Sensor Fusion for Efficient Autonomous Occupancy Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for generating free-space and occupancy maps using image-based approaches or LiDAR/RADAR data are less reliable due to inaccuracies and require significant computational resources, making them inefficient for autonomous and semi-autonomous navigation.
Innovation Solution
A sensor fusion-based method using ultrasonic sensors in conjunction with other sensors like RADAR, LiDAR, and image sensors, processed by neural networks to generate accurate height and occupancy maps, reducing computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If image-based approaches are used to generate free-space and occupancy maps, then the system can process visual data, but the reliability of distance and depth determinations deteriorates
Solution Approach 1:
The patent combines multiple sensor types (ultrasonic, RADAR, LiDAR, image sensors) into a unified sensor fusion system. This merging allows the system to leverage the complementary strengths of each sensor type, where ultrasonic sensors provide reliable proximity data for near-field objects, RADAR offers robust mid-range detection, and LiDAR/cameras contribute visual information, thereby resolving the contradiction between visual data processing capability and measurement reliability.
Solution Approach 2:
The patent creates a composite sensing system that integrates data from heterogeneous sensor modalities. Similar to how composite materials combine different materials to achieve superior properties, this composite sensor system combines the distance-measurement strengths of ultrasonic, RADAR, and LiDAR sensors with the visual processing capabilities of image sensors, achieving both versatility and reliability simultaneously.
2Measurement precision
If LiDAR is used to generate heatmaps and occupancy maps, then measurement accuracy improves, but computational requirements and expense increase
Solution Approach 1:
The patent segments the sensing space into different ranges and assigns appropriate sensor types to each segment. Ultrasonic sensors handle near-field proximity detection, RADAR covers mid-range detection, and LiDAR is used selectively for long-range or specific critical measurements. This segmentation allows the system to achieve high measurement precision where needed while reducing overall computational complexity and sensor expense by not deploying LiDAR across the entire sensing volume.
Solution Approach 2:
The patent applies LiDAR processing partially rather than comprehensively to all sensor data. Instead of processing all LiDAR point clouds through computationally intensive heatmap generation, the system uses LiDAR data selectively for specific occupancy map regions or critical object detections, thereby maintaining measurement precision for essential functions while reducing overall computational requirements.
3Loss of information
If RADAR or LiDAR data is processed alone to determine object locations, then depth information can be obtained, but noise and errors in processing reduce map reliability
Solution Approach 1:
The patent merges data from multiple sensor types to compensate for the noise and errors inherent in individual sensor processing. By combining ultrasonic proximity data, RADAR velocity and range information, and LiDAR spatial mapping, the system creates redundant measurement paths that cross-validate each other, thereby maintaining depth information availability while significantly improving map generation reliability through error cancellation and noise reduction.
4Ease of manufacture
If conventional image-based approaches are used for map generation, then the system can operate with standard sensors, but the computational efficiency deteriorates
Solution Approach 1:
The patent replaces computationally intensive image-based processing with a hybrid sensor fusion approach that uses dedicated depth-sensing hardware (ultrasonic, RADAR, LiDAR) for direct distance and occupancy measurements. This substitution of mechanical/optical measurement systems for computational image processing methods maintains ease of manufacture through standard sensor availability while dramatically improving map generation productivity by reducing computational complexity.
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
The method provides more accurate and efficient map generation with reduced computational resources, enhancing navigation capabilities for autonomous systems.
Implementation Method 1
sensor data generated using one or more ultrasonic sensors
Data Source
AI summary
In various examples, techniques for sensor-fusion based object detection and/or free-space detection using ultrasonic sensors are described. Systems may receive sensor data generated using one or more types of sensors of a machine. In some examples, the systems may then process at least a portion of the sensor data to generate input data, where the input data represents one or more locations of one or more objects within an environment. The systems may then input at least a portion of the sensor data and/or at least a portion of the input data into one or more neural networks that are trained to output one or more maps or other output representations associated with the environment. In some examples, the map(s) may include a height, an occupancy, and/or height/occupancy map generated, e.g., from a birds-eye-view perspective. The machine may use these outputs to perform one or more operations.