Robot Sensor Layout for Blind-Spot-Free Multi-Sensor Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing robot sensing system has many blind areas and low cooperation degree among sensors, leading to low accuracy in multi-sensor fusion algorithms and reduced robustness, which affects the robot's ability to autonomously perceive the environment and perform tasks effectively.

Innovation Solution

A robot sensor arrangement system is designed with fixed positions for image sensors and inertial sensors relative to each other, ensuring they do not move with external conditions like vibration and temperature changes, and strategically positioned at specific angles to enhance obstacle avoidance and system robustness. The system includes a combination of image sensors, inertial sensors, optical ranging sensors, mechanical odometers, and ultrasonic sensors, each responsible for its own collection scope, improving data fusion and cooperation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If multiple sensors are arranged on the robot body to enable autonomous perception, then the robot can perform tasks autonomously, but the sensor assembly creates many blind areas and low cooperation degree

Engineering Contradiction:
Improveautonomous perception capabilityVSAvoidsensing system robustness
Core Design Contradiction:
Extent of automationVSReliability

Solution Approach 1:

The sensor system is divided into multiple sensor assemblies, each with specific functional responsibilities. Image sensors capture visual information while inertial sensors measure motion parameters, creating segmented sensing zones that collectively cover the robot's environment without blind spots.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Multiple sensor types (image sensors and inertial sensors) are merged into integrated sensor assemblies with fixed relative positions. This combination enables coordinated data collection where visual and motion data are fused to improve autonomous perception reliability and eliminate blind areas.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If multiple sensors collect the same physical quantity simultaneously, then redundancy is provided, but the cooperation degree is low and fusion algorithm accuracy is reduced

Engineering Contradiction:
Improvedata redundancyVSAvoidfusion algorithm accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

Different sensor assemblies are assigned different measurement responsibilities based on their spatial locations and capabilities. Image sensors focus on visual feature extraction while inertial sensors focus on motion parameter measurement, creating localized measurement quality optimization that improves overall fusion accuracy.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The sensor system dynamically assigns measurement tasks based on operational context. When the robot is stationary, image sensors perform detailed environmental scanning; when moving, inertial sensors provide continuous motion tracking. This dynamic division of labor improves cooperation degree and fusion precision.

Inventive Principle:
Principle #15Dynamics

3Reliability

If sensor positions are fixed relative to each other, then sensor cooperation is improved, but the system becomes sensitive to external conditions like vibration and temperature

Engineering Contradiction:
Improvesensor cooperation stabilityVSAvoidexternal environmental sensitivity
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

Inertial sensors provide continuous feedback on the robot's motion state and orientation. This feedback is used to dynamically compensate for position drift in image sensors caused by vibration and temperature changes, maintaining accurate spatial relationships between sensor data despite external disturbances.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3885077B1Robot sensor arrangement system
Publication Date: 2024.08.14 JUXING TECH SHENZHEN CO LTD
  • EP3885077B1 patent drawingFigure 1

AI summary

A robot sensor arrangement system. At least one sensor assembly is arranged on a robot body (20), wherein the sensor assembly comprises image sensors (1001, 1002) and a first inertial sensor (1007), and the positions of the image sensors (1001, 1002) relative to the first inertial sensor (1007) are fixed such that the image sensors and the first inertial sensor (1007) do not move as external physical conditions, such as vibration and temperature change. The included angle between the positions of the image sensors (1001, 1002) and a vertical axis is in a first angle range so as to ensure the robot can autonomously sense the surrounding environment to improve the capability of autonomous obstacle avoidance and the robustness of a robot system.