Longbin's Tech-Blog

唐隆斌

工作经历

感知算法工程师

小马智行 Pony.ai 2025.7 - 至今

负责自动驾驶感知算法的研发与迭代，参与多传感器感知系统中的目标检测、跟踪与场景理解模块优化。围绕真实道路数据进行问题分析、数据闭环、模型训练与效果评估，提升复杂交通场景下感知系统的鲁棒性与稳定性。

技术：自动驾驶感知，多传感器融合，目标检测/跟踪，Python，C++，深度学习

感知算法实习生

星海图 GalaXeaAI 2024.2 - 2024.6

负责无人车感知数据的生成、处理以及方案设计与优化。参与基于 LiDAR 监督的 3D Occupancy 感知模型训练，同时协助进行路径规划算法（基于深度学习）的测试与优化工作，协助机械臂控制算法调试与遥操作开发。

技术：3D Occupancy，SLAM，ROS，Python，Docker

项目和研究经历

Agile Flight Traversing Dynamic Environment With Learning Based Planning and Control

新加坡国立大学 2024.9 - 2025.4

通过深度强化学习和可微分 MPC，实现 FPV 无人机在高速状态下穿越多个动态物体，并安全到达目标点。在Isaac Sim中搭建自定义无人机仿真平台，并部署 PX4 后端以实现高效且真实的并行训练，确保优秀的 Sim2Real 迁移效果。最终在室内动捕环境中完成实际飞行测试，成功实现无人机在不同初始状态下穿越动态环境的任务。

技术：深度强化学习、MPC、Isaac Sim、无人机 (UAV)

Autonomous Vehicle for mail and parcel delivery

新加坡国立大学（苏州）研究院 & 西交利物浦大学 2023.10 - 2024.6

设计一辆能够在园区（例如学校、公园）内的自主导航与配送的无人车。底盘采用六轮差速结构，配备双叉臂独立悬架，实现零转弯半径与稳定配送。采用 Autoware.universe 架构，“Lidar-IMU-Odometer-GPS-Radar”多传感器实现 SLAM、点云分割、路径规划与运动控制，并利用深度强化学习实现更适合园区复杂三维路面的局部路径规划。

技术：SLAM，ROS2，C/C++，Python，LQR/MPC，Reinforcement Learning

论文发表

Kuankuan Sima, Longbin Tang, Zhenyu Yang, Haozhe Ma, Lin Zhao. MacroNav: Multi-Task Context Representation Learning Enables Efficient Navigation in Unknown Environments. RA-L 2026.
Xin Chen, Rui Huang, Longbin Tang, Lin Zhao. AERO-MPPI: Anchor-Guided Ensemble Trajectory Optimization for Agile Mapless Drone Navigation. ICRA 2026.
Tianchen Sun, Bingheng Wang, Longbin Tang, Yichao Gao, Lin Zhao. Learning Agile Gate Traversal via Analytical Optimal Policy Gradient. IROS 2026.
Runyu Lei, Longbin Tang, Jiaxin Guo, Jie Sun, Qinglei Bu. Autonomous Delivery Vehicle System Based on Multi-Sensor Data Fusion. IRAC 2024.

专利成果

唐隆斌, 王锐, 何铭, 刘能峰, 一种差分舵轮及小车, 公开号 CN117104334A
唐隆斌, 郭佳鑫, 雷润宇, 卜庆雷, 孙捷, 一种无人驾驶车辆的车轮控制方法以及装置

Longbin Tang

Work Experience

Perception Algorithm Engineer

Pony.ai 2025.7 - Now

Develop autonomous driving perception algorithms, focusing on object detection, tracking, and scene understanding in multi-sensor perception systems. Work with real-world driving data for issue analysis, data-loop iteration, model training, and evaluation to improve perception robustness and stability in complex traffic scenarios.

Techs: Autonomous Driving Perception, Multi-Sensor Fusion, Object Detection/Tracking, Python, C++, Deep Learning

Perception Algorithm Intern

GalaXeaAI 2024.2 - 2024.6

Handled autonomous vehicle perception data generation, processing, and optimization. Worked on 3D Occupancy perception models supervised by LiDAR, tested and optimized deep learning-based path planning algorithms, and assisted with robotic arm control and teleoperation development.

Techs: 3D Occupancy, SLAM, ROS, Python, Docker

Projects & Research

Agile Flight Traversing Dynamic Environment With Learning Based Planning and Control

National University of Singapore 2024.9 - 2025.4

“Enabled an FPV drone to navigate dynamic environments and reach targets safely using deep reinforcement learning and differentiable MPC. Developed a custom simulation platform in Isaac Sim with PX4 backend for efficient training and strong Sim2Real transfer, validated through successful indoor flight tests.”

Techs: Reinforcement Learning, MPC, Isaac Sim, UAV

Autonomous Vehicle for Mail and Parcel Delivery

NUSRI & XJTLU 2023.10 - 2024.6

“Designed an autonomous vehicle for campus navigation and delivery using a six-wheel differential chassis with double-wishbone suspension for zero turning radius and stable delivery. Integrated ‘Lidar-IMU-Odometer-GPS-Radar’ sensors with Autoware.universe for SLAM, point cloud segmentation, path planning, and motion control. Employed deep reinforcement learning for local path planning on complex campus surfaces.”

Techs: SLAM, ROS2, C/C++, Python, LQR/MPC, Reinforcement Learning