Robot navigation with stereo vision and artificial intelligence

Subject description:

Robot navigation in outdoor environments is a problem investigated by several researchers from the area of: computer science, robotics, computer vision, artificial intelligence, control theory and automation. This is due to the fact that robot navigation involves problems of: trajectory planning, path following, control strategy, perception and filtering. Nowadays with the exponential popularity of drones and autonomous vehicles, such a problem is in vogue in the industry and in several research institutes all around the world. This thesis proposal aims to contribute in the solution of the autonomy problem of mobile robots.

During this thesis the student must acquire knowledge in the following areas and expertise:

• Probability
• Mathematical analysis
• Nonlinear systems
• Theory of neural networks
• Python
• Robot Operating System
• OpenCV
• Keras
• Tensor Flow
• Git
• Probabilistic robotics

This is a challenging thesis including science, technology and even (as a function of the student performance) the possibility of innovation. It is recommended that the student has some expertise in coding, math and developing engineering projects.

The thesis is developed in collaborations with PhD, master and undergraduate students who are Lapyr LAB members.

Objectives

• Design an algorithm for autonomous navigation in outdoor environments making use of artificial and vision intelligence.
• Publish a paper in a prestigious international conference such as: ICRA, IROS, ACC, CVPR, ICUAS.
• Attending the conference.
• A draft for an International Journal.
• The thesis

State of the project

Project started in 2017.

Students involved in the project:

• Daniela Martínez (CIO)
• Sergio Trejo(CIO)
• Luis Arreola(CIO)
• Gesem Gudiño(CIO-UTL)

Publications from the team:

• Improvement in the UAV position estimation with low-cost GPS, INS and vision-based system: Application to a quadrotor UAV. L. Arreola; A. Montes de Oca ; A. Flores ; J. Sanchez ; G. Flores. 2018 International Conference on Unmanned Aircraft Systems (ICUAS); pp.1248 – 1254; 2018.
• Other works in preparation from students.

Materials

Stereo cameras; monocular cameras; Jetson TX2; quadrotor drone; mobile robot; autopilots; Lidar Velodyne; altitude sensors; gyro sensors. All material is available at the LAB.

Financial support

Consorcio de Inteligencia Artificial. The first autonomous vehicle in Mexico in collaboration with the Free University of Berlin and CIMAT.