|CMPE-633A/CS-633: Robot Motion Planning|
Dr Abubakr Muhammad, Assistant Professor of Electrical Engineering
Email: abubakr [at] lums.edu.pk
Office: Room 9-309A, 3rd Floor, SSE Bldg
Office Hours: Mon, Wed: 1215-1300; Tue, Thurs: 1130-1300 or by appointment.
Teaching assistant. Talha Manzoor. Room 9-309. talha.manzoor [at] lums.edu.pk
Elective course for electrical engineering, computer engineering and computer science majors
Course Website: http://cyphynets.lums.edu.pk/index.php/CMPE-633
A research-methods based course to study advanced topics in robotics and control system design with emphasis on field robotics, unmanned aerial and ground vehicles, planning algorithms, autonomous systems, telerobotics, Human Robot Interaction (HRI) and other related areas. The course prepares students to do independent work at the frontiers of robotics and control research.
Objectives (this year!)
- Introduce fundamental principles in robot motion planning.
- Use of geometric and dynamical models acquired from sensory data.
- Using sensor-based information to determine robot’s own state and of the world
- Control theoretic issues in trajectory planning and sensory feedback.
- Introduce practical applications
This course is NOT about
- Mechatronics or robot building.
- Higher-level perception and AI.
Courses. CMPE432. Feedback control systems OR CMPE435. Robotics OR By permission of instructor.
Co-requisites. CMPE-501. Applied probability OR by permission of instructor.
Topics/Skills. Programming proficiency in C or MATLAB; multi-variable calculus, linear algebra, probability
The course will be taught from a combination of the following textbooks.
- Principles of Robot Motion by Choset et al. [available as low priced edition]
- Planning Algorithms by Steve Lavalle. [available as a free legal download]
- Probabilistic Robotics by Thrun et al.
- Research papers.
- Class Participation: 15%
- Assignments: 15%
- Final Exam: 20%
- Project: 50%
- Proposal. 10%
- Report. 20%
- Presentation. 20%
- Code. 50%
Course Delivery Method
Lectures. Mon, Wed: 10:30am-11:30am. 10-402. SSE Bldg.
|Week 1. August 23||Aug 23. Classes begin.|| Lec 1. Introduction; robotics and autonomous systems;
Lec 2. workspaces; configuration spaces; planning algorithms; bug algorithms; Bug0 and Bug1 algorithms;
|Choset CH 1,2;|
|Week 2. August 30||Aug 30. Add/drop with full refund|| Lec 3. Completeness of Bug1; upper bounds on Bug1; Bug2 algorithm; performance comparison; range sensors and mathematical description.
Lec 4. Tangent Bug algorithm; implementation issues; wall-following behavior with range sensors;
|Choset CH 2; LaValle SEC 12.3.3;|
|Week 3. September 6||Sept 9-14. Eid-ul-Fitr holiday|| Lec 5. Introduction to discrete planning; state space models and examples; discrete feasible planning; graph search algorithms; general forward search;
Lec 6. Breadth first search; depth first search; Dijkstra's algorithm; A*; systematic searches; configuration spaces revisited; mechanical linkages; forward and inverse kinematics; toroidal configuration spaces;
|Week 4. September 13||Sept 15. Second payment deadline||Lec 7. Mapping workspace obstacles into configuration space obstacles; visualizing high dimensional configuration spaces; a first look at obstacles in SE(2); polygonal robots and obstacles;||Choset CH3, Appendix F|
|Week 5. September 20|| Lec 8. representing polygonal objects via linear inequalities; computing configuration space obstacle polygons from workspace descriptions; star algorithm; Minkowski sums and differences of sets;
Lec 9. Holonomic constraints; degrees of freedom in a configuration space; rigid body CSpace in 2D; rigid bodies in 3D; SO(3) matrix group; Euler parameterization;
| Lavalle SEC4.3.2, SEC3.2;
Choset CH3, Appendex F, E.
|Week 6. September 27||October 1. Semester Drop with fee Penalty|| Lec 10. SO(2) revisited via complex numbers; generalization to SO(3); quaternion algebra; topology of SO(3);
Lec 11. Gimbal lock problem; parameterization problems and topology of configuration spaces; topology of S^1 x S^1 vs S^2; S^3; SO(3); RP^3; quaternions as a double cover of SO(3);
| Lavalle SEC3.2 ,SEC4.2; Choset CH3, Appendix E;
Gimbal lock problem in Euler Angles. Video.
|Week 7. October 4|| Lec 12. 2D and 3D rigid bodies revisited; rotations and translations of points and bodies; types of joints; kinematic chains; DH-parameters;
Lec 13. Introduction to roadmaps; deterministic/combinatorial methods for roadmap construction; general properties; visibility graphs; line-sweep algorithm;
|LaValle SEC3.3; Choset CH5.|
|Week 8. October 11|| Lec 14. Voronoi diagrams; generalized Voronoi diagrams (GVD); bush-fire algorithm; wavefront algorithm;
Lec 15. Potential field methods; vector fields and gradients; critical points and Hessian matrix; attractive and repulsive potential fields; gradient descent algorithm; local minima and other practical implementation issues; another look at bush-fire algorithm;
|Week 9. October 18||Midterm exams|| Lec 16. Introduction to navigation functions; navigation functions for the sphere-shaped world; compositions of potentials; analytical switches; empirical tuning of navigation functions;
Lec 17. Star-shaped worlds; diffeomorphisms; mapping stars to spheres; analytical switches revisited; navigation functions for star worlds;
|Week 10. October 25|| Lec 17. Inverse and forward kinematics; lifting velocities and forces between coordinate changes; modified potential field method for direct workspace control;
Lec 18. Introduction to sampling based planning; problem history and motivation; computational complexity of planning algorithms; introduction to probabilistic road-maps algorithm (PRM); query phase of PRM;
Tutorial. Using the MATLAB robotics toolbox. (Oct 29)
|Choset SEC3.8, SEC4.7, CH 7;|
|Week 12. November 1|| Project Proposal Presentations. (Session 1)
Project Proposal Presentations. (Session 2)
|Week 13. November 8||Nov 9. Iqbal day;|| Lec 20. PRMs continued. Refinements in PRM construction; sampling difficulties and narrow passages; distance functions for non-Euclidean spaces; metric spaces and Lp-norms for complex configuration spaces; Cartesian products; metrics for S^1, SO(3) using complex numbers and quaternions; generating a random configuration in S^1, SO(3);
Lec 21. Introduction to probabilistic robotics; modeling sensor noise; basic probability and statistics; linear transformations of Gaussian random variables; interpreting and visualizing the covariance matrix; state and measurement;
Lec 22. Linear systems; discrete-time linear models from equations of motion; examples; process noise and sensor noise; Objectives of Kalman filtering;
|Choset CH7, CH8;|
|Week 14. November 15||Eid-ul-Azha holidays. Nov 17-19.|
|Week 15. November 22||
Lec 23. Basic laws of probability and Bayes rule; belief about state; derivation of the Bayesian filter; prediction and update; dealing with beliefs and distributions for nonlinear non-Gaussian systems; Kalman Filtering as a special case;
Lec 24. Derivation of the Kalman filter; combining estimates from two scalar Gaussians; innovation and Kalman gain; estimation in the absence of process and sensor noise; state and covariance prediction; combining measurement and prediction in the presence of process noise;
| Thrun CH 2; Choset CH8
|Week 16. Novemeber 29||
Lec 25. Completion of proof of KF; basic SLAM using Kalman filtering; non-linear models of sensing and robot motion; Extended Kalman filtering;
Lec 26. Nonlinear transformations of random variables; generation of non-Gaussian noise via nonlinear transforms; modeling sensor noise; ultrasound, laser models and other non-Gaussian sources; data association problem;
Lec 27. Probabilistic localization using Bayesian filtering; data association problem revisited; histogram Bayes filter; introduction to particle filtering; computational issues in sampling arbitrary distributions; practical demonstration of Kalman filtering.
|Thrun CH 2,3,4; Choset CH8|
|Week 17. December 6||Dec 8. Last day of classes; Dec 9-15. Final Exams||
Lec 27. Wrap-up. Future directions.
Take home Exam. Due December 7th.
|Week 18. December 13|| Dec 16-17. Ashura Holidays;
Dec 27. Final grades submission; Dec 20-Jan 21 Semester break.
| Project Presentations/Viva/Demo 1. Dec 14th
Project Presentations/Viva/Demo 2. Dec 15th
Feedback control methods / robot dynamics
- LQR-Trees: Feedback Motion Planning on Randomized Trees. Paper..
- LQG-MP: Optimized Path Planning for Robots with Motion Uncertainty and Imperfect State Information Paper. Hasan Arshad Nasir, 10060010.
- Integrated Planning and Control for Convex-bodied Nonholonomic systems using Local Feedback Control Policies Paper.
- Gait Regulation and Feedback on a Robotic Climbing Hexapod Paper.
- Unified Inverse Depth Parametrization for Monocular SLAM Paper.
- Monte-Carlo Localization for Mobile Robots with Stereo Vision Paper.
- Probabilistic Terrain Analysis For High-Speed Desert Driving Paper.
- A Bayesian Approach to Nonlinear Parameter Identification for Rigid Body Dynamics Paper.
- The Iterated Sigma Point Kalman Filter with Applications to Long Range Stereo Paper. Ateeq ur Rehman Shaheen, 09060002.
- Gaussian Processes for Signal Strength-Based Location Estimation Paper. Abdul Rehman Aslam, 09060014.
- Map-Based Precision Vehicle Localization in Urban Environments Paper. Muhammad Salman, 09060016.
- Detection of Principal Directions in Unknown Environments for Autonomous Navigation Paper. Amer Zaheer, 04030068.
- Improving Localization Robustness in Monocular SLAM Using a High-Speed Camera Paper.
- Mapping Large Loops with a Single Hand-Held Camera Paper. Rehmat Ullah Khattak, 08060009.
- Model Based Vehicle Tracking for Autonomous Driving in Urban Environments Paper. Mazher Ahmad, 10060014.
- Large Scale Graph-based SLAM using Aerial Images as Prior Information Paper. Syed Muhammad Abbas, 10030019.
Sampling based methods / Roadmaps
- Incremental Sampling-based Algorithms for Optimal Motion Planning Paper. Ibrahim Tariq, 10060016.
- Roadmap Based Pursuit-Evasion and Collision Avoidance Paper. Mudassir Khan, 10060013.
- Elastic Roadmaps: Globally Task-Consistent Motion for Autonomous Mobile Manipulation in Dynamic Environments Paper. Mhequb Hayat, 09060019.
- Structural Improvement Filtering Strategy for PRM Paper.
- Single-Cluster Spectral Graph Partitioning for Robotics Applications Paper.
- Dynamic Maps for Long-Term Operation of Mobile Service Robots Paper.
- Distributed Coverage Control with Sensory Feedback for Networked Robots Paper. Muhammad Irfan Riaz, 09060005.
- Dubins Traveling Salesperson Problems: novel approximation algorithms Paper.
- Composition of Vector Fields for Multi-Robot Manipulation via Caging Paper. Nauman Shahid, 10060025.
Robotics Technology Trends
[An evolving compilation!]
- A robot in every home by Bill Gates (Scientific American Magazine, 2007).
- A Robot in Every Home by 2020, South Korea Says (National Geographic Magazine, 2006)
- Kiva Systems Link.
- iRobot Link.
- Boston Dynamics Link.
- KUKA Link.
- Willow Garage Link.
- Robotics at Honda [Link.]
- Robotics at Kawasaki [Link.]
- Robotics at Seiko Epson [Link.]
- Robotics at Sony [Link.]
- Robotics at TOSY [Link.]
New Issues of Ethics and Law in Robotics
- International Committee for Robot Arms Control Link.
- Isaac Asimov's Three Laws of Robotics.
- Human Robot Interaction HRI
- United States Air Force Unmanned Aircraft Systems Flight Plan 2009-2047 Report.
- Unmanned Systems Roadmap 2007-2032 Report.
- Report to US Congress: Development and Utilization of Robotics and Unmanned Ground Vehicles Report.
- Autonomous Vehicles in Support of Naval Operations Report.