Quan Khanh Luu

Pengyuan Guo, Zhonghao Mai, Zhengtong Xu et al.

Recent advances in vision-language models (VLMs) have enabled increasing progress in real-world robot manipulation. However, long-horizon manipulation in unstructured environments requires VLMs to reason about changing scene states, action constraints, and execution outcomes, which remains difficult with natural language reasoning alone. We present PLanAR, a planning-language-grounded robot agent framework for open-vocabulary, long-horizon manipulation. PLanAR uses a planning-language interface to define the VLM reasoning space: object predicates represent scene states, action schemas specify robot skills with preconditions and effects, and symbolic plans provide executable intermediate representations. This interface enables stepwise verification: after each action, PLanAR uses onboard observations to check whether the expected symbolic effects have been achieved, allowing the VLM-based agent to update task states, detect failures, and replan when execution deviates from expectation. Across robot embodiments, VLM backends, and tasks including stacking, crossword solving, and long-horizon kitchen workflows, PLanAR demonstrates strong real-world capability while revealing key limitations of current VLMs in embodied reasoning.

Valerio Belcamino, Nhat Minh Dinh Le, Quan Khanh Luu et al.

Human activity recognition (HAR) is essential for effective Human-Robot Collaboration (HRC), enabling robots to interpret and respond to human actions. This study evaluates the ability of a vision-based tactile sensor to classify 15 activities, comparing its performance to an IMU-based data glove. Additionally, we propose a multi-modal framework combining tactile and motion data to leverage their complementary strengths. We examined three approaches: motion-based classification (MBC) using IMU data, tactile-based classification (TBC) with single or dual video streams, and multi-modal classification (MMC) integrating both. Offline validation on segmented datasets assessed each configuration's accuracy under controlled conditions, while online validation on continuous action sequences tested online performance. Results showed the multi-modal approach consistently outperformed single-modality methods, highlighting the potential of integrating tactile and motion sensing to enhance HAR systems for collaborative robotics.

Son Tien Bui, Quan Khanh Luu, Dinh Quang Nguyen et al.

There is a growing need for vertical take-off and landing vehicles, including drones, which are safe to use and can adapt to collisions. The risks of damage by collision, to humans, obstacles in the environment, and drones themselves, are significant. This has prompted a search into nature for a highly resilient structure that can inform a design of propellers to reduce those risks and enhance safety. Inspired by the flexibility and resilience of dragonfly wings, we propose a novel design for a biomimetic drone propeller called Tombo propeller. Here, we report on the design and fabrication process of this biomimetic propeller that can accommodate collisions and recover quickly, while maintaining sufficient thrust force to hover and fly. We describe the development of an aerodynamic model and experiments conducted to investigate performance characteristics for various configurations of the propeller morphology, and related properties, such as generated thrust force, thrust force deviation, collision force, recovery time, lift-to-drag ratio, and noise. Finally, we design and showcase a control strategy for a drone equipped with Tombo propellers that collides in mid-air with an obstacle and recovers from collision continuing flying. The results show that the maximum collision force generated by the proposed Tombo propeller is less than two-thirds that of a traditional rigid propeller, which suggests the concrete possibility to employ deformable propellers for drones flying in a cluttered environment. This research can contribute to morphological design of flying vehicles for agile and resilient performance.