Tin Lai

About

Tin Lai is a robotics researcher whose work sits at the intersection of motion planning, probabilistic inference, and autonomous navigation. His most significant contributions center on improving the efficiency of sampling-based motion planners — a dominant paradigm in robotics — by replacing uninformative global sampling distributions with intelligent, locally adaptive strategies. His 2020 paper on Bayesian Local Sampling-Based Planning (43 citations) introduced a probabilistic framework that dramatically reduces wasted samples in complex, constrained environments, while subsequent work on PlannerFlows leveraged normalising flows to learn high-quality motion samplers from data. Lai has also made notable contributions to Visual-SLAM, co-authoring a widely cited 2022 review (32 citations) synthesising advances from geometric modelling to learning-based semantic scene understanding through multi-modal sensor fusion. His research extends into human-robot coexistence, including anticipatory crowd navigation, and more recently into trajectory optimisation using path signatures and adaptive movement primitives. Across his career, Lai has accumulated over 120 citations, reflecting growing recognition of his efforts to make robot motion planning smarter, more data-driven, and better suited to the complexities of real-world environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Bayesian Local Sampling-Based Planning

   43 citations · 2020
2. 2
   A Review on Visual-SLAM: Advancements from Geometric Modelling to Learning-Based Semantic Scene Understanding Using Multi-Modal Sensor Fusion

   32 citations · 2022
3. 3
   Adaptively Exploits Local Structure With Generalised Multi-Trees Motion Planning

   16 citations · 2021
4. 4
   PlannerFlows: Learning Motion Samplers with Normalising Flows

   11 citations · 2021
5. 5
   Learning to Plan Optimally with Flow-based Motion Planner

   5 citations · 2020
6. 6
   Anticipatory Navigation in Crowds by Probabilistic Prediction of Pedestrian Future Movements

   4 citations · 2021
7. 7
   Rapidly-exploring Random Forest: Adaptively Exploits Local Structure with Generalised Multi-Trees Motion Planning

   4 citations · 2021
8. 8
   Path signatures for diversity in probabilistic trajectory optimisation

   3 citations · 2024
9. 9
   Stein Movement Primitives for Adaptive Multi-Modal Trajectory Generation

   2 citations · 2024
10. 10
    Local Sampling-based Planning with Sequential Bayesian Updates.

    2 citations · 2019

Key Collaborators

Contact & Links