Ke Ma Email & Phone Number

Shanghai, China

• Vector Mapping: An onboard road element mapping and topology reconstruction system. Implemented 1) lane centre line smooth algorithm based on Composite Bˊezier Curves fitting and curvature optimisation using gradient descend approach with the ability of obstacle avoidance; 2) intersection detection; 3) traffic lights to lane topology bindings; 4) grid-based lane estimation via traffic flow in intersection area; 5) road markers clustering and mapping from multiple observations; 6) map… Show more • Vector Mapping: An onboard road element mapping and topology reconstruction system. Implemented 1) lane centre line smooth algorithm based on Composite Bˊezier Curves fitting and curvature optimisation using gradient descend approach with the ability of obstacle avoidance; 2) intersection detection; 3) traffic lights to lane topology bindings; 4) grid-based lane estimation via traffic flow in intersection area; 5) road markers clustering and mapping from multiple observations; 6) map management in SQLite database.• Crowdsource vector mapping: Implemented road net (nodes and skeletons) construction algorithm, node mapping and topology reconstruction via map alignment and map fusion. Show less

Pudong, Shanghai, China

• Lidar-Camera extrinsics calibration: Implementated an onboard 3D lidar to Ladybug 360°camera calibration algorithm via multiple flat chessboards for HDMap collection vehicles, and constructed a MySQL database with calibration parameters served for those vehicles.• Map Change Detection: Led a cross-functional team in implementing an end-to-end map change detection system.The full system includes onboard lane boundary markings and traffic lights change detection, data packaging and… Show more • Lidar-Camera extrinsics calibration: Implementated an onboard 3D lidar to Ladybug 360°camera calibration algorithm via multiple flat chessboards for HDMap collection vehicles, and constructed a MySQL database with calibration parameters served for those vehicles.• Map Change Detection: Led a cross-functional team in implementing an end-to-end map change detection system.The full system includes onboard lane boundary markings and traffic lights change detection, data packaging and uploading, detailed map change detection via various data source from multiple vehicles, and risk map building for autonomous vehicles and HDMap repair.• Risk Map Cloud Service: Implemented a cloud based micro-service that collects map change area from the previous system (via Kafka message) and adds into a risk map (database) as a complementary layer for HDMap requested (via Kafka messages) by cloud navigation system to end users. Show less

Shanghai, China

• Semantic localisation: Implemented an EKF based localisation with sensor fusion, using camera, IMU/GNSS and wheel odometry. My contribution is on semantic elements alignment with two approaches – alignments on image plane and ground plane separately. Average horizontal error less than 20cm.• Lidar point cloud localisation: Two cloud alignment approaches: 1) Normal distribution transform (NDT) and extended to NDTi with additional intensity dimension added; 2) Image based alignment by… Show more • Semantic localisation: Implemented an EKF based localisation with sensor fusion, using camera, IMU/GNSS and wheel odometry. My contribution is on semantic elements alignment with two approaches – alignments on image plane and ground plane separately. Average horizontal error less than 20cm.• Lidar point cloud localisation: Two cloud alignment approaches: 1) Normal distribution transform (NDT) and extended to NDTi with additional intensity dimension added; 2) Image based alignment by projecting cloud onto ground plane, and scan alignment based on three map layers, i.e., altitude map, intensity map and pole map.• Lidar-camera extrinsics calibration: Implemented a calibration algorithm that estimate lidar to camera extrinsic parameters with reprojection error of 5 pixels in 1920x1080 image. Show less

Guangdong, China

Led the SLAM team developing 2D SLAM algorithms for construction field AGVs, based on open sourced algorithms, such as GMapping and AMCL.

Sheffield, United Kingdom

My main role is to provide guidance and support in laboratories, mark assignment and exam papers for undergraduate and postgraduate students in the department. Teaching modules include: Foundations of Control Systems, Introduction to Aerospace Design, Advanced Industrial Control, Intelligent Systems.

Doctor Of Philosophy (Ph.D.), Robot Navigation

My project is focused on robot simultaneous localisation and mapping (SLAM) in water pipes in the Assessing The Underworld (ATU) project.

Master Of Science (Msc), Control Systems, Distinction

Faculty Student Representative & Department Student Representative Foundations of Control Systems, State Space Modelling, Digital Signal.

Bachelor'S Degree, Electrical And Electronics Engineering

Exchange Scheme Award: Outstanding Academic Achievement the OPENFX.ORG Prize, 2012.