Jingxi Xu AI Research

Improve human-robot interaction and rehabilitation through advanced intent detection, tactile sensing, and synthetic data generation for assistive tools.

Free robotics research

About Jingxi Xu AI Research

Jingxi Xu’s AI research portfolio provides a collection of machine learning frameworks and robotic control systems designed to facilitate interaction between human intent and robotic hardware. The primary focus of the work is the development of assistive and rehabilitation technologies, such as robotic hand orthoses that aid stroke patients in regaining motor control. These systems utilize multimodal biosignals, specifically surface electromyography (sEMG), to predict user intended movement in real-time. By integrating generative AI techniques, the research addresses the challenge of data scarcity in medical robotics, allowing for the creation of robust control models that can adapt to individual user needs without requiring large amounts of manually collected data. The portfolio features several notable projects, including ChatEMG and GEOTACT. ChatEMG utilizes synthetic data generation to improve the control performance of robotic hand orthoses, enabling manipulation and feedback for patients with neurological conditions. GEOTACT focuses on tactile-based object retrieval, allowing robots to identify and extract buried objects within granular media using touch sensors alone. These tools utilize algorithms for active tactile exploration and zero-shot intent detection, exploring how robots perceive and interact with environments where visual information is limited. The research also encompasses visual navigation, multi-arm motion planning, and optimal control. This work is intended for researchers in robotics, biomedical engineering, and human-computer interaction. Healthcare providers and occupational therapists can utilize the insights and wearable technologies developed for stroke recovery, while roboticists can apply the tactile sensing and manipulation policies to industrial or domestic tasks. The tools are useful for scenarios where data is limited or where interaction with hidden objects is necessary. The research specifically focuses on "sparsity and scarcity," providing resources for developers working with small datasets or complex sensing requirements. This research differs from standard robotics software through its integration of clinical needs with machine learning. The tools are specifically designed to handle the variability of human biosignals and the physical requirements of rehabilitation. By combining tactile intelligence with intent inferral, the research establishes a framework for human-robot collaboration. Additionally, the availability of codebases for projects like ChatEMG and ReactEMG on GitHub provides a transparent and accessible foundation for the broader robotics community.

Jingxi Xu AI Research pros & cons

Pros

Utilizes synthetic data generation to overcome data scarcity in rehabilitation robotics.
Supports zero-shot, low-latency intent detection for immediate user interaction.
Provides tactile sensing policies that allow for object retrieval in complex granular media.
Research is validated through high-tier peer-reviewed publications and awards.
Offers open-source access to several codebases and datasets via GitHub.

Cons

Most tools are research-oriented and require technical expertise to implement.
Full functionality often depends on specific hardware like robotic orthoses or tactile sensors.
Real-world medical application requires professional clinical supervision.
Documentation is primarily focused on academic audiences rather than end-users.

Jingxi Xu AI Research use cases

Researchers can utilize the GEOTACT framework to develop robots capable of identifying objects through touch in obstructed environments.
Stroke patients can benefit from the MyHand orthosis which uses AI to translate muscle signals into hand movements.
Bioengineers can use the ChatEMG codebase to generate synthetic sEMG data for training more robust gesture recognition models.

Jingxi Xu AI Research features

panoramic visual navigation
multi-arm motion planning
robotic hand orthosis control
zero-shot model adaptation
active exploration policies
tactile object retrieval
synthetic biosignal generation
surface emg intent detection

Jingxi Xu AI Research pricing

Is Jingxi Xu AI Research free? Yes, Jingxi Xu AI Research has a free plan.

Research Access

Free

Open-source code
Pre-trained models
Synthetic datasets
Hardware specifications
Academic publications

Jingxi Xu AI Research FAQs

What is ChatEMG?

ChatEMG is an AI framework that creates synthetic electromyography data to improve the control systems of robotic hand orthoses. This addresses the challenge of insufficient training data available from stroke patients who cannot perform extensive movement trials.

Can the robotic hand orthosis detect user intent in real-time?

Yes, the ReactEMG project focuses on zero-shot, low-latency intent detection via biosignals. This allows the robotic system to predict and execute the user's desired movements immediately without lengthy calibration sessions.

How does GEOTACT use tactile sensing?

GEOTACT implements tactile exploration policies that allow robots to find and retrieve objects buried in materials like sand or grain. It relies on touch sensors to recognize object geometry when visual sensors are occluded.

Are the datasets and codebases accessible for external use?

Many of the research projects provide links to public GitHub repositories and datasets. You can find these links alongside the respective research publications listed on the portfolio website for collaborative or academic use.

Ratings & reviews

No reviews yet. Be the first to share how Jingxi Xu AI Research worked for you.