Hybrid Lidar Camera Sensor for SLAM Error Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing sensor systems for mobile robots face challenges in converging data from Lidar and camera sensors, which are different in nature, leading to difficulties in generating accurate maps and predicting positions in unknown environments, especially as errors accumulate with increased travel distance and number of nodes in SLAM techniques.

Innovation Solution

A hybrid sensor module that maps distance information from a Lidar sensor to image information from a camera sensor, interpolating distance information based on pixel intensity to generate composite data, which is then used to create a broader area map by generating key frames, calculating odometry and loop closure edges, and updating positions to predict global maps.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If data from Lidar sensor and camera are used separately, then each sensor can operate independently, but the data cannot be converged effectively leading to incomplete environmental recognition

Engineering Contradiction:
Improvedata convergenceVSAvoidsensor integration
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent combines Lidar and camera sensors into a hybrid sensor system that captures both depth information from Lidar and image information from camera simultaneously. The processing unit integrates these different data types to generate composite data, merging the capabilities of both sensors to achieve complete environmental recognition without losing information from either source.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The processing unit acts as an intermediary that receives data from both Lidar and camera sensors, processes and converges them into composite data. This intermediary component enables effective data convergence by transforming and integrating the different data types from the two sensors into a unified format that can be used for comprehensive environmental mapping.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If graph based SLAM is used to represent position and motion, then the robot can track its movement, but errors accumulate when traveling distance increases and number of nodes increases

Engineering Contradiction:
Improveposition prediction accuracyVSAvoidtraveling distance
Core Design Contradiction:
Measurement precisionVSDuration of action of moving object

Solution Approach 1:

The hybrid sensor system provides continuous feedback by capturing both depth and image information throughout the robot's movement. This dual-sensor feedback mechanism allows for more accurate odometry calculations and loop closure detection, correcting position errors as they accumulate during long-term operation and preventing drift in position prediction over extended traveling distances.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent creates a composite sensing system that combines Lidar depth data with camera image data, analogous to composite materials combining different properties. This composite data approach integrates the complementary strengths of both sensors to maintain measurement precision over long durations, preventing error accumulation that would occur with single-sensor systems.

Inventive Principle:
Principle #40Composite materials

3Area of stationary object

If map generation is performed using sensor information, then the robot can navigate unknown environments, but the map coverage area is limited by sensor capabilities

Engineering Contradiction:
Improvemap coverage areaVSAvoidmap accuracy
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The hybrid sensor merges Lidar's depth measurement capability with camera's high-resolution image capture capability. This combination allows the system to generate maps that cover broader areas while maintaining high accuracy, as the Lidar provides spatial structure information and the camera provides detailed visual information, together creating comprehensive and accurate environmental representations.

Inventive Principle:
Principle #5Merging (Combining)

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 hybrid sensor module effectively converges Lidar and camera data, enabling the generation of more accurate and extensive maps with reduced error accumulation, improving the robot's ability to navigate and map unknown environments.

Implementation Method 1

A time of flight based Lidar measures a time of a light signal which is emitted and then reflected and measures a distance from a reflector using a speed of light

Methodology Applied
Scientific EffectTime of flight: Time of Flight

Data Source

PatentEP3447532B1Hybrid sensor with camera and lidar, and moving object
Publication Date: 2024.09.11 YUJIN ROBOT
  • EP3447532B1 patent drawingFigure 1~2
  • EP3447532B1 patent drawingFigure 3~4
  • EP3447532B1 patent drawingFigure 5

AI summary

The present exemplary embodiments provide a hybrid sensor and a moving object which generate composite data by mapping distance information on an obstacle obtained through a Lidar sensor to image information on an obstacle obtained through an image sensor and predict distance information of composite data based on intensity information of a pixel, to generate precise composite data.