Intelligent Mobile Robotics

Learning Outcomes

Learning Outcomes

Having successfully completed this module you will be able to:

• Understand the key issues surrounding uncertainty of sensors, actuators and dynamic models used in different robotic application domains that motivate the need for probabilistic methods. [SM-M1, EA-M2]
• Understand the use of statistical modelling techniques (e.g., Gaussian Processes) to allow robots to interpret sensor data and make sense of their surroundings. [SM-M1, EA-M2]
• Understand the operating principles, assumptions and limiting factors of different Bayesian frameworks (e.g., Kalman Filters, Particle Filters, Graphical Methods) in the context of robot localisation and mapping. [SM-M1, EA-M2, EA-M3]
• Understand how probabilistic methods can address the uncertainty that is inherent due to real-world non-determinism. [SM-M1, EA-M2]
• Be able to design practical experiments in order to assess the success and sensitivities of implemented methods in practical problem solving tasks. [EA-M2, EA-M3, DI-C4, EP-C12]
• Be able to adapt and apply robotic concepts to design and develop practical robotic solutions for different application domains. [EA-C4, EP-C13]
• Be able to accurately express basic robotic concepts and critically assess their suitability for different application domains based on their assumptions and their limitations. [EA-M2, EA-M3, EA-C4, EP-C17]
• Knowledge of how to implement probabilistic methods on a simple mobile robot using python language and robot middleware (e.g., ROS). [EA-C3, DI-C6, EP-C12]
• Understand the principles of dynamic modelling, sensing and closed-loop control used in mobile robotics. [SM-M1, EA-M2]
• Knowledge of social, environmental impacts of robotics and automation, including ethical considerations, security risks, diversity and inclusion in future job markets. [ES-C7, ES-C8, ES-C9, ES-C11]

The module has two series of lectures and associated practicals. Lectures teach the theory and highlight areas of additional reading. Practicals implement methods taught in the lectures on a small robotic platform. Practical sessions are in small groups, and consist of laboratory based tutorials and design workshops that provide each group with opportunities for data gathering, performance demonstration and presentation. Learning outcomes are assessed through two assignments, each with equally weighted group and individual components.

Part A: The Robot: Probabilistic Localisation

Lectures:

• Introduction to probabilistic robotics
• Fundamentals of kinematics and control
• Fundamentals of sensing
• Bayesian framework for recursive state estimation
• Probabilistic localisation with Extended Kalman Filters
• Probabilistic localisation with Particle Filters

Practicals:

• Programming and platform basics (tutorial and workshops)
• Kinematics and control (tutorial and workshops)
• Probabilistic localisation with Extended Kalman Filters (tutorial and workshops)
• Probabilistic localisation with Particle Filters (tutorial and workshops)

Assignment A: The Robot - Probabilistic Localisation: Students implement probabilistic localisation methods on a small autonomous robotic platform to conduct a mission using a combination of motion sensors and a dynamic motion model. Each group to provides a live group demonstrate/presentation and experimental results generated form the basis of an individual report to demonstrate learning outcomes.

Part B: The World: Perception, Localisation and Mapping

Lectures:

• Introduction to Simultaneous Localisation and Mapping (SLAM)
• Perception and reasoning
• Gaussian Processes for interpreting sensor observations
• Feature-based SLAM part 1
• Feature-based SLAM part 2
• Feature-based SLAM part 3
• Featureless SLAM part 1
• Featureless SLAM part 2
• Featureless SLAM part 3

Practicals:

• Feature-based SLAM using pose graphs (tutorial and workshops)
• Featureless SLAM using particle trajectories (tutorial and workshops)

Assignment B: The World: Perception, Localisation and Mapping: Students implement perception and SLAM frameworks for real-time operation of a small autonomous robotic platform to conduct mission using observations of its surroundings. Each group to provides a live group demonstrate/presentation and experimental results generated form the basis of an individual report to demonstrate learning outcomes.

Teaching and learning methods

Teaching methods include

• ~6 lectures on robotics fundamentals and probabilistic localisation
• ~9 lectures on perception and reasoning for localisation and mapping
• ~8 tutorials on robot programming and implementation of methods
• ~10 workshop experimental sessions focused on the two assignments

Learning activities include

• Directed reading/independent learning
• Directed use of robotic software libraries
• Implementation of probabilistic methods on provided robotic hardware in tutorials and assignments
• Design of experimental methods to demonstrate learning outcomes for the two assignments
• Task management when working in groups for the two assignments
• Report-writing for the two assignments
• Live-demonstration and presentation for the two assignments

Assessment strategy

Referral Method: There are two assignments in this module. If the mark achieved for the module is less than the module pass mark then a referral will be required.

The referral will be an individual assignment, where the student will be provided with experimental data on which to perform analysis and produce a written report.

(Summative) Assessment Method

Assignment – The Robot: Probabilistic localisation 40 %Feedback uploaded to blackboard site after each assignment is marked.

Assignment - The World: Perception, Localisation and Mapping 60 %Feedback uploaded to blackboard site after each assignment is marked.

(Referral) Assessment MethodAssignment -Assignment(s) 100 %

Method of repeat year:

Repeat year externallyYes

Repeat year internallyYes

Formative

This is how we’ll give you feedback as you are learning. It is not a formal test or exam.

Summative

This is how we’ll formally assess what you have learned in this module.

Referral

This is how we’ll assess you if you don’t meet the criteria to pass this module.