Sensing, Control, and Analysis of Networked and Multi-Agent Systems

IMG_0518-1Next-generation engineering systems are expected to provide service far beyond the capability of a single machine. As a consequence, many individuals will have to work collaboratively to achieve an advanced task. However, challenges arise in the case where many agents interact with each other. Moreover, the agent could be either a machine or a human subject, which makes the decision making and control system design very complicated. In this project, fundamental challenges of designing such systems are investigated. We envision this research will provide methodologies for designing various engineering systems, including unmanned aerial vehicles (UAVs), autonomous ground vehicles, and multi-robot cooperative teams.

High-level perception and decision making

  • Modeling of interactions between agents
  • Decision making and motion/behavior planning of agents
  • Safety, robustness, and resilience of multi-agent systems

Low-level networked motion control

  • Network scheduling
  • Time delay/packet loss compensation
  • Precision motion control of individual agent

Student Researchers: Shatadal Mishra, Dangli Yang, YiZhuang Garrard, Brandon Dawson

Gait Analysis and Control of Assistive Robot for NeuroRehabilitation

  • Smart Shoes: Human Gait Monitoring System Based on Air Pressure Sensors Embedded in Shoes

Smart-Shoes-page-001-300x179A wireless gait monitoring system is designed for the diagnosis of abnormal gait and rehabilitation. Measurement of ground contact forces (GCFs) provide necessary information to detect gait phases. The silicone tubes are wound into air bladders connect to four barometric pressure sensors to measure GCFs on the toe, the first metatarsophalangeal joint (Meta1), the fourth metatarsophalangeal joint (Meta4), and the heel. We are planning to develop a human activity recognition (HAR) system using smart shoes and inertial measurement units (IMUs). Also, HAR system will be integrated with knee joint exoskeleton to provide useful feedback about the activities such as walking on level ground, going upstairs and going downstairs.

Student Researchers: Prudhvi Tej Chinimilli, Zhi Qiao, Seanwolfgang Wachtel, Breanna Hassett, Julie Vuong, Wenhao Deng 

  • Wearable Robotics: Knee Joint Exoskeleton
Isometric-with-white-bkgrd-web-218x300Knee Joint Exoskeleton is one part of a robotic suit for human physical augment and gait rehabilitation. It assists patients with motor disabilities (i.e. stroke, spinal cord injury and Parkinson’s disease, etc.) in regaining ability to stand and walk, and also help therapists during their patients’ rehabilitation training. The Series Elastic Actuator (SEA) used in this exoskeleton has high maximum force/speed, excellent position controllability and back-drivability. It offers excellent torque regulation, light weight as well as compactness. An intelligent controller will be implemented with Electromyography (EMG) sensors as well as a adaptive control algorithm for walking on flat ground, upstairs and downstairs.

Student Researchers: Prudhvi Tej Chinimilli, Mostafa Rezayat, Vaibhav Jhawar, Zhi Qiao, Wenhao Deng

Intelligent Control and Optimization for Human-robot Collaboration

Nowadays, humans work in close cooperation with robots more than ever, 555which makes safety an important consideration in physical human-robot interactions (pHRI). Robots can perform powerful movements that can be dangerous to the human co-workers. Almost all of the industrial robots are manufactured under certain safety standard, and this limits the operation speed of the robot and requires every robot to have protective stop function. Fatal accidents still happen when human workers and the robot manipulators are operating in the same working cell. To ensure the safety of human robot coexisting working space, robots should be perceptive to the human activity and conduct safe motion plan respect the human’s intention. This research is focus on the algorithms of human activity perception, safety evaluation, real-time collision free trajectory generation and the pHRI models, which are present in a more cooperative way.

Student Researchers: Yiwei Wang, Zhiang Chen


Comments are closed