CS计算机代考程序代写 matlab algorithm SCHOOL OF ELECTRICAL AND ELECRONIC ENGINEERING

SCHOOL OF ELECTRICAL AND ELECRONIC ENGINEERING

Electrical & Electronic Engineering, Software & Electronic Systems Engineering
Final Year Projects 2020-2021

Planning offloading in mobile robotics

Supervisor: Dr Nikolaos Athanasopoulos Moderator: Dr Mien Van

X
Control

Embedded Systems

High Frequency Electronics

Microelectronics

Electric Power
X
Software

Connected Health

MEMS

Cyber-Security

Wireless Communications
X

Signal/Image Processing
X
Intelligent Systems

Digital Design

Sensor Networks
X
Data Analytics

Electronics

Modern mobile robots are supposed to find their way, i.e., plan paths, avoid obstacles, move in dynamic environments. This has been recently possible by the advancement of sensing devices, cheap hardware and new smart decision mechanisms. However, there are many tasks to be carried out that are computation-intensive and memory-intensive. On top of that these tasks are often safety- and time-critical, meaning they should be carried out in real time. One solution is to use simpler algorithms running locally on a microcontroller, however this is not always enough as complicated algorithms have to be ran to ensure safety, especially in path planning and localization and mapping. Another solution is to utilize the cloud, effectively offloading all computation intensive tasks, with the risk of creating congestion in the communication network, or violating timing contracts when there is delay.

This project will focus on the problem of deciding when to offload tasks related to path planning of mobile robots (and inevitably localization tasks as well) . A specific path planning algorithm will be analysed and simulated in a simulation environment (e.g., using ROS, MATLAB or other software of preference). Additionally, a modelling of the communication network and the remote (edge or cloud) computing infrastructure will be made. Finally, a new path planning offloading mechanism will be developed that guarantees the robot will reach a predetermined goal without violating constraints, and at the same time utilize in an optimal way the remote resources and network.

Objectives
1. Familiarize with basic planning algorithms for mobile robots (simple differential/unicycle dynamics).
2. Learn about localization and mapping algorithms.
3. Program in a simulation environment the robot dynamics, the environment, and the planning algorithms together with the trajectory tracking controller (using e.g., Matlab, ROS).
4. Propose a simple offloading model of the available resources in cloud/edge and integrate it with the closed-loop mobile robot model.
MEng Extension

1. Analyse stability/convergence/safety properties of the offloading mechanism.

Learning Outcomes
Upon completion of the project you will expect to:
1. Learn about planning methods in mobile robots.
2. Learn about dynamical models of computation/memory/communication resources in a control engineering application/
3. Gain experience in programming and simulation
4. Learn about a significant control problem in robotics and in general in cyber-physical control systems.
5. Develop good mathematical and algorithmic reasoning.