Hybrid Lidar Camera Sensor for SLAM Error Reduction
Find Innovative SolutionsGenerate 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
Engineering 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
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.
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.
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
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.
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.
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
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.
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
Data Source
Figure 1~2
Figure 3~4
Figure 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.