LiDAR Development Project
What led to the introduction of the service?
The client looked into the possibility of implementing object recognition and autonomous mobile robots by remote sensing using Light Detection and Ranging (LiDAR). They also wanted to develop communication between LiDAR modules and in-vehicle devices for a separate project, as well as mobility with autonomous driving technology and driver assistance systems.
Issues before introduction
The client had little experience in LiDAR development and lacked knowledge in related fields such as remote sensing, autonomous driving, object recognition, and Simultaneous Localization and Mapping (SLAM). Another internal issue was the lack of development personnel available for new development. Therefore, considering the opportunity for the client to accumulate technologies, they request us, as we have extensive experience in LiDAR-related technologies, to develop the system.
Examples of use after introduction
We successfully developed a LiDAR-based mobility system by installing LiDAR on existing vehicles. We also carried out implementation and unit evaluation of an autonomous mobile robot by applying LiDAR-based self-localization estimation to SLAM.
Effectiveness and achieved results
We also successfully solved the client’s issues such as lack of development experience in LiDAR and lack of knowledge in related fields, and were able to provide the client with mobility solutions using LiDAR.
・Specification review process
Conducted examinations on driver assistance system specifications and constructed simulated environments.
Conducted everything from implementation to evaluation of individual units for autonomous mobile robots aimed at 3D SLAM with the goal of accumulating in-house technologies.
Created a tool to evaluate the validity of recognition results by sensors and give feedback for parameter adjustment.
Consideration process for driver assistance system specifications
・Consideration of the concept of an object recognition function Integrated the concept of recognition principles and algorithms into the specifications. *Mathematical knowledge is required.
Created a pseudo environment for Ethernet communication using CANoe.
・2 microcomputer control
Examined asynchronous processing of data control and communication control microcomputers.
Implementation process for autonomous mobile robots
・Building a development platform
Built a robotic environment which acts as a development platform for autonomous mobility byprocuring commercially available development-based robots and LiDAR. Prepared power supply, wireless environment, development laptops, and more.
・SLAM function implemented
Applied the open-source library gmapping and Google's cartographer after installing the ROS environment to achieve SLAM (Simultaneous Localization and Mapping). In cartographer, the environment is able to be seen in 3D and autonomous movement can be programmed.
Evaluation process for object recognition algorithm
・Recognition algorithm (developed by another company)
The system detects objects based on LiDAR sensor data and outputs them as tagged recognition results.
・Evaluation and analysis tools (Developed by us)
In addition to calculating the percentage of correct answers by linking the recognition results from the recognition algorithm with the correct answer data, the system can output and display information such as what objects were not recognized with just two software configuration.