• GitHub
• Link

1. MXT, London
   1. Stonehenge VR Visualisation
      1. Environment Loading
      2. Traffic
      3. Artist Tools
      4. Animal AI
   2. Driving Simulator
      1. The Road Editor
      2. Real-time Signs and Signals
      3. Vehicle Simulation
2. Personal
   1. XOR Neural Network in C++
   2. Ez-Utils
3. Jam Games
   1. Xanadu
   2. Fall
   3. Numb Skull
   4. HADRON
   5. Chimney
   6. Another Planet

MXT, London

Stonehenge VR Visualisation

This was our immersive experience covering the 22km² world heritage site around the Stonehenge monument. It contained a free roam and guided tour experience of the environment built using real-world data. This sandbox experience created in Unreal allowed the users to explore the environment in a number of configurations. While standing anywhere in the environment, you can instantly move into the future to view the landscape after the proposed tunnel development. There was also a seasonal configuration for a midday in summer and during the winter solstice.

Environment Loading

One aspect of my work on this project was building a system for handling the loading of environment. With such a significant map size, which requires both long distance view between the monuments of the landscape, while also requiring instant changes in configurations of the environment, the environment provided a challenge (especially in VR).

I developed a system using the World Composition of Unreal Engine 4 to dynamically load a developer configured list of landscape tiles and levels while culling unnecessary tiles for performance. This allowed us to develop the two landscapes for the current and future scenarios.

Traffic

The environment loading system also linked to the traffic system which is present in both time scenarios. As the project was data-driven the traffic was a visualization of measured real-world traffic along the A303. I created the pipeline for taking the PT Vissim traffic, which was two-dimensional, and parsing it into Unreal Engine and matching the vehicles to the virtual environment. Then I created the system for visualizing the over 1200 vehicles, leveraging C++ and Hierarchical Instanced StaticMeshes to maintain performance on VR headsets.

Artist Tools

With two times of day and two environments, one for each time period, there was a great deal of overhead for artist. I created a number of tools to improve their workflows. These included an editor buttons to easily move between times of day while working and scripted actions to allow artists to duplicate their objects and changes between landscapes or easily move objects between them.

Animal AI

A small detail (and personal passion of mine) that was added to the environment was animals. When at Stonehenge in real life you will regularly see birds and livestock around. I designed and created AI behaviour for the Sheep, Cows, Crows and Starlings present around the environment. I used Unreal Engine’s Behaviour Tree and NavMesh tools to manage the movements, behaviour and animations of the grazing animals.

However, the starlings required more custom behaviour. They needed to fly in a murmuration and go between a state of flying and sitting atop the stones. Therefore I created a simple system (similar to the start of a primative boid system) which allowed a leader to start a murmuration and fly smoothly between a range of predefined points before returning to their resting places on the stones. Each of the “follower” starlings will attempt to keep within a specified distance of the leader while flying in a smooth manner.

---

Driving Simulator

This project is a simulator with an immersive driving experience created within Unity involving a Mixed Reality experience with a headset and motion rig. Alongside this driver experience is a real time traffic simulation using SUMO. This setup, along with our analytics suite, allows us to study driver behaviour and biometric data in a broad range of scenarios.

The Road Editor

The core of the immersive environment created for the simulations is a tool for creating realistic roads which are integrated with the SUMO traffic simulation workflow.

This is a large existing C# codebase for a Unity editor tool for generating objects and meshes procedurally while alowing for it to exported into a network to run with SUMO simulations. I am the lead programmer responsible for the development and maintaining of this tool, working closely with other programmers and the artists using it. While working on this tool I have implemented a number of features including:

An action history and undo system, utilising command pattern.

Serializing the network to a custom XML .mxt file type, made easy for developers with custom attributes.

Efficient traffic routing using Breadth-First Search algorithm.

Real-time Signs and Signals

To allow for great diversity in scenarios I was responsible for creating the system managing runtime variable traffic signals which affect the SUMO traffic simulation.

This required adding to the network code used in communication between the Unity project and the SUMO traffic simulation (written in Python). These modifications would allow the changes in the complex signage to be sent to the other connected clients and traffic rules to the simulation.

Vehicle Simulation

Another area of my work on this project was in the accurate simulation of the feel of driving. I worked on creating accurate driving feel and force feedback through the motion platform based on the simulated driver vehicle rigidbody.

This system involved connecting the Unity simulation with the ForceSeatMI API. Additionally I created safety features to register when the simulated rigidbody’s motion, wheel contact or collisions were unsuitable for a driver to experience and stop motion matching between the two systems.

---

Personal

XOR Neural Network in C++

---

Ez-Utils

---

Jam Games

---

---

---

---

---