Khashayar Hojjati Emami presented last Friday

Khashayar Hojjati Emami presented his PhD thesis seminar work last Friday.  The title of the talk was “Human Centered Reliability Assessment and Condition Monitoring of Risks in Road Transportation Systems”.

DSC_9203-2

Advertisements

TA training this Friday!

TA training seminar this Friday!  This seminar is MANDATORY for all new TAs and for all those hoping to become a new TA next semester.  Experienced TAs are also invited to come share their experience with the newer TAs.

Date: Friday September 19th

Time: 2:30pm

Room: CBY B205

Hope to see you there!

Cameron Frazier presented his MASc thesis seminar on Friday

This past Friday, Cameron Frazier presented his MASc thesis seminar, with a talk titled “Re-Active Vector Equilibrium (RAVE): A Novel Method of Autonomous Rover Local Navigation Using Potential Fields.”  The abstract of the seminar is below.  Check out the photo of Cameron with his happy supervisor! 😉

 

Abstract

The use of potential eld based navigation schemes in robotics has been limited by inherent local minima issues. Local minima traps, small passages, unstable motion, and targets positioned near objects all pose major concerns when using potential fields for local vehicle control. This work proposes a new algorithm, “Re-Active Vector Equilibrium” (RAVE) that mitigates many of these issues. The vehicle representation model is expanded to use multiple points and the addition of two forces, a velocity dependent risk force and a velocity and direction dependent tangential force. Expanding the vehicle representation model from a single reactive point to a series of points that define the vehicle body is also done, providing better and simpler vehicle control. This has the effect of simplifying the required calculations at the cost of increasing the calculation count. The risk force allows for dynamic adaptation to the immediate environment by acting in opposition to the net obstacle force, and is
inversely proportional to the vehicle speed. The tangential force encourages better wall-following behaviour and provides a biasing mechanism to resolve obstacle aligned with target local minima issues.  Presented here is a brief background on the topic, a description of the proposed algorithm, presentation of simulations and results, and presentation of implementation videos.

 

Cameron and supervisor (Dr. Baddour)

Cameron and supervisor (Dr. Baddour)