Dr David Howard

Honorary Fellow

School of Information Technology and Electrical Engineering
Faculty of Engineering, Architecture and Information Technology

Overview

Research Interests

  • Evolutionary computing
  • Neural networks
  • Robotics
  • Shape optimisation
  • UAV control

Available Projects

  • I am seeking a student to manage the evolutionary optimisation of legged vehicle morphology in simulation. The project aims to discover the best body structures for legged vehicles with an outlook to informing the design of some of our future legged platforms. The role will involve optimisation of morphology, trialling different approaches and body types. It will require the academic and software expertise to:

    - identify the best morphological representation (the genome)

    - generate and improve the suite of locomotion experiments

    - identify and improve fitness metrics

    - improve the evolutionary algorithm if needed

    - monitor and identify simulation problems

    As such it is beneficial for the candidate to have experience with:

    - C++

    - physics simulations

    - robotics

    - optimization, evolutionary algorithms preferred

  • Develop an interface between the Solidworks CAD package and the chrono:engine simulator that lets us easily print components that have just been simulated and tested. Develop a common definition for this capability that can be globally used to bridge the gap between CAD packages and simulators. Can also look to extending across multiple simulators.

    The goal of this project is to look at how to represent different types of geometries, in such a way that they can be modified by a Machine Learning algorithm. So we automatically design shapes and test in the simulator, but can also physically print them easily as they’re CAD files. We need the pipeline that lets us do this.

  • Searching for a student with interests in robotic simulators and code optimisation - the project is to modify the Voxcad simualtion package to allow for larger-scale robots to be simulated. Voxcad models soft robots as a series of contiguous deformable voxels. Each voxel can be one of several types, e.g., contracting muscle, expanding muscle, soft tissue, hard (bone) support.

    There are two main approaches to be investigated:

    1) Split a large robot deterministically into a number of equal sections, simulate each section individually and create methods that allow the sections to effectively communicate their deformations to each other

    2) Analyse the composition of a large robot, and create large continuous sections of a single type,which can be replaced with a larger-scale voxel of the same type.

    We can also have a look at parallelising the simulator onto a high-performance cluster.

View all Available Projects

Possible Research Projects

Note for students: The possible research projects listed on this page may not be comprehensive or up to date. Always feel free to contact the staff for more information, and also with your own research ideas.

  • I am seeking a student to manage the evolutionary optimisation of legged vehicle morphology in simulation. The project aims to discover the best body structures for legged vehicles with an outlook to informing the design of some of our future legged platforms. The role will involve optimisation of morphology, trialling different approaches and body types. It will require the academic and software expertise to:

    - identify the best morphological representation (the genome)

    - generate and improve the suite of locomotion experiments

    - identify and improve fitness metrics

    - improve the evolutionary algorithm if needed

    - monitor and identify simulation problems

    As such it is beneficial for the candidate to have experience with:

    - C++

    - physics simulations

    - robotics

    - optimization, evolutionary algorithms preferred

  • Develop an interface between the Solidworks CAD package and the chrono:engine simulator that lets us easily print components that have just been simulated and tested. Develop a common definition for this capability that can be globally used to bridge the gap between CAD packages and simulators. Can also look to extending across multiple simulators.

    The goal of this project is to look at how to represent different types of geometries, in such a way that they can be modified by a Machine Learning algorithm. So we automatically design shapes and test in the simulator, but can also physically print them easily as they’re CAD files. We need the pipeline that lets us do this.

  • Searching for a student with interests in robotic simulators and code optimisation - the project is to modify the Voxcad simualtion package to allow for larger-scale robots to be simulated. Voxcad models soft robots as a series of contiguous deformable voxels. Each voxel can be one of several types, e.g., contracting muscle, expanding muscle, soft tissue, hard (bone) support.

    There are two main approaches to be investigated:

    1) Split a large robot deterministically into a number of equal sections, simulate each section individually and create methods that allow the sections to effectively communicate their deformations to each other

    2) Analyse the composition of a large robot, and create large continuous sections of a single type,which can be replaced with a larger-scale voxel of the same type.

    We can also have a look at parallelising the simulator onto a high-performance cluster.