Welcome to Robot Intelligence Laboratory

Vision and Mission

Intelligent robots have been envisioned to weave a hybrid society with human beings for years, but the coming of this robot revolution still seems unclear in the very near future. One of the underlying reasons is that robots have not yet got the ability to understand the world and human instructions. The vision of the Robotic Intelligence Laboratory is to endow an intelligent robot with the ability to autonomously accomplish navigation and manipulation tasks by interpreting human instructions and perceiving the environment. The robots with such an ability will autonomously assemble components into final products by following an instructional manual. The key to achieve the goal is a trustworthy architecture that covers robot perception, decision making, and human-machine interaction. In the architecture, humans and robots understand and share knowledge through a human-machine “common language”. Different from artificial intelligence. Robust analytic methodology will be developed for modeling, analysis and verification of intelligent robots, environments, and human instructions.

Research Thrusts

Smart Perception

Smart perception is the foundation of intelligent robotics. One research direction is smart sensing technologies for real-time environmental sensing and perception, including scene contents and dynamics. In particular, intelligent processing algorithms for heterogeneous sensor fusion will be studied for depth perception, motion perception, 3D object detection, and object recognition.

Robot Understanding

Another research direction is a formal cognitive framework that enables an intelligent robot to understand sensor inputs and human instructions. A unified representation of natural-language instructions and sensor perception will be studied so that ambiguous human instructions and unstructured environmental information can be converted into structured semantic expressions. The structured semantic expressions will be used in robot reasoning and learning.

Automata

The third research direction is autonomous action planing for different types of robots to complete navigation or manipulation tasks. Human instructions and environmental perception will be grounded to Abstract Robot Actions in accordance with the underlying logic and dependencies across the whole instruction discourse. The Abstract Robot Actions provide an interface for various robots with different perception and operation capabilities. The presentations, action reasoning, and action planning are verified using formal methods with respect to the logic and constraints in the instructions.

Recent News

  • GRASP 2018 Award Winners: 1st Place
    Bharath Rao, Department of Industrial, Systems & Manufacturing Engineering and Department of Electrical Engineering & Computer Science. Advisors: Dr. Hongsheng He and Dr. Krishna Krishnan. Topic: "Learning Robotic Grasping Strategy Based on Natural-Language Object Descriptions."

Hongsheng He, PhD

Assistant Professor,
Director of Robot Intelligence Laboratory,
Department of Electrical Engineering and Computer Science,
Wichita State University

CV (PDF)
Google Scholar

Address

Robot Intelligence Laboratory
330 Wallace Hall, Wichita State University
1845 Fairmount St, Wichita KS USA 67260-0083
Location on Google Map

Contacts

Office: 210 Jabara Hall
Tel: (316) 978-3053
Mail: Campus Mailbox 83
Email Me