Objectvie of Job
Embedded in a worldwide network Mercedes-Benz Research & Development China continuously strives to remain at the forefront of successful automotive research and development. Mercedes-Benz R&D China is headquartered in Beijing and is opening a R&D branch office in Shanghai, with key areas of Autonomous Driving(AD), Adavanced Driving Asistant System (ADAS), Cloud Platform for ADAS/AD, Advanced Interaction Design, Telematics, Big Data Analytics, and Open Innovation. Your key results include
Our AD AI Algo team is looking for a highly motivated and skilled Engineer to support with our perception development. At this role you will have the unique opportunity to work with a group of innovative engineers together to improve driving safety and customer driving experience.
Task Description
• Design and develop MB next generation 4D imaging radar AI perception algorithm, include advanced radar object, road boundary detection DNN, tracking and post processing algorithms.
• Support EOL calibration algorithm development, include lidar, camera, lidar-camera joint calibration algorithm development and validation.
• Analyze current calibration target, and implement targeted algorithm improvements to enhance EOL calibration accuracy and robustness.
• Develop robust online calibration algorithms for lidar to estimate high-precision extrinsics in dynamic scenarios.
• Responsible for radar perception algorithm software implementation and deployment in AD ECU.
• Develop and maintain radar perception related data pipeline and model pipeline for continued data driven development.
• Cross function coordinate between radar hardware, radar service, radar perception algorithm and customer function owner, to ensure radar perception can meet the requirement of related AD system functions.
Qualification
• Education background in Computer Science/Robotics/Electronics/Automation or other related majors.
• Deep understanding and project experience on radar perception or any point cloud perception classic algorithms and neural network algorithms.
• Proficient in lidar point cloud processing, noise removal, and feature extraction methods; skilled in nonlinear optimization techniques (e.g., Ceres, GTSAM); familiar with mainstream lidar SLAM approaches.
• Proficient in image feature extraction and multi-view geometry principles; experienced with at least one static calibration method (e.g., checkerboard calibration).
• Solid research background on 3D object detection, object tracking, robotics and multiple sensor calibration such as camera, lidar, IMU.
• Project experience with data pipeline and model pipeline in data-driven development.
• Proficient in C++ and Python programming. Good code style is needed.
• Familiar with ROS programming and related tools usage.
• Experience in EOL calibration algorithms for mass-produced vehicle models is a plus.