Responsible for the design, implementation, and testing for 3D mapping algorithms, for autonomous drones.
● Design advanced 3D mapping algorithms and systems for autonomous drones. The sensor modalities include camera, Lidar, Radar, IMU, GPS, and more.
● Research and evaluate different candidate algorithms to determine the best option based on computational requirements of the system, accuracy of the results, and applicability of the algorithm to the specific use case.
● Document all feature design.
● Implement these advanced algorithms in sophisticated computer programming languages, such as C++, and deploy the program onto different robotic drone platforms.
● Perform rigorous testing to ensure implemented code meets all functional requirements.
● Master‘s degree or above;
● Familiar with classic dense SLAM methods, such as DynamicFusion, ElasticFusion, and BundleFusion.
● Familiar with classic map representation and update methods, such as voxel map (e.g., ESDF, TSDF, occupancy, and NDT), mesh model, etc.
● Deeply understand the multi-view geometry of the camera. Have a solid foundation in probability theory, linear algebra and space geometry, and have a deep understanding of various matrix factorization algorithms, linear equations solutions, Lie algebra/groups, and spatial coordinate transformations in different coordinate systems.
● Familiar with C++/Python language, familiar with the use of optimized libraries such as Ceres, g2o, etc.
● Experience with NeRF.