SENG402 Abstracts - Software Engineering Showcase | Whakaaturanga Hanga Pūmanawa 2016 - Computer Science and Software Engineering - University of Canterbury - New Zealand

SENG402 Abstracts 2016

James Harrison

UAV Swarm: Localisation within a known indoor environment utilising fiducial markers

During the operation of UAVs inside an indoor environment, it can prove difficult to determine their precise location. GPS signals can be blocked by both external and internal walls and therefore cannot be relied upon for accurate positioning. This paper proposes an approach to utilise fiducial markers and an RGB camera for accurate localisation within a known indoor environment. The research was conducted as a part of a team project with the goal of having an autonomous swarm of UAVs collaboratively map a room while maintaining video links to a base station.

Bradley Kirwan

Multi-hop ad hoc communication over WiFi in a UAV Swarm

UAV Swarms flying autonomously present many challenges for communication. In some cases, there will not exist a direct communication channel between each UAV and a ground control station, therefore multi-hop communication is required. A system has been designed to select the optimal route from a node (UAV) to the sink (ground control station), based on the cost associated with each link. Further, the system makes use of IP Source Routing to stream live video to the sink, with an emphasis on low-latency. This research is a part of a team project enabling a swarm of UAVs to collaboratively map an environment with the future goal of navigating forests to prune trees.

Sam Schofield

UAV Swarm: Autonomous Environment Mapping

A system for autonomously mapping a 3D environment has many potential applications including disaster zone evaluation, terrain mapping, and building inspection. This paper presents an autonomous mapping system, which can be easily integrated with existing camera based simultaneous localisation and mapping (SLAM) algorithms. The proposed solution was developed as part of a larger team project, aiming to map an environment autonomously using a swarm of UAVs, while maintaining video links to a base station.

James Fairbairn

Telogis Job Location Inference Detection (team project)

The aim of this project is to infer job locations from vehicle point data. This project is done in partnership with Telogis. Telogis provides fleet organisation and route optimisation, Software as a Service (SaaS) solutions. Jobs are visits to a site by a vehicle, examples of this are visits by a truck to a distribution centre or a courier delivering packages to a building. Currently these jobs are manually entered into the system. The objective of this project is to generate a list of jobs from vehicle point data, with a lower number of inaccuracies. That is few false positives and no false negatives. The rationale behind the project is to reduce the number of jobs which are entered manually. This could save customers of Telogis time and also allow customers that were not manually entering jobs to use the features that they enable. Although this project has been worked on by a team of students, each have specialised in a particular area. Matthew will be discussing data labelling and clustering of arbitrary point data. James will cover the machine learning methods used to form the prediction models. Dion will be demoing the reporting interface used to display the results of the project and change variables of the algorithms used.

Matthew Knox

Telogis Job Location Inference Detection (team project)

The aim of this project is to infer job locations from vehicle point data. This project is done in partnership with Telogis. Telogis provides fleet organisation and route optimisation, Software as a Service (SaaS) solutions. Jobs are visits to a site by a vehicle, examples of this are visits by a truck to a distribution centre or a courier delivering packages to a building. Currently these jobs are manually entered into the system. The objective of this project is to generate a list of jobs from vehicle point data, with a lower number of inaccuracies. That is few false positives and no false negatives. The rationale behind the project is to reduce the number of jobs which are entered manually. This could save customers of Telogis time and also allow customers that were not manually entering jobs to use the features that they enable. Although this project has been worked on by a team of students, each have specialised in a particular area. Matthew will be discussing data labelling and clustering of arbitrary point data. James will cover the machine learning methods used to form the prediction models. Dion will be demoing the reporting interface used to display the results of the project and change variables of the algorithms used.

Dion Woolley

Telogis Job Location Inference Detection (team project)

The aim of this project is to infer job locations from vehicle point data. This project is done in partnership with Telogis. Telogis provides fleet organisation and route optimisation, Software as a Service (SaaS) solutions. Jobs are visits to a site by a vehicle, examples of this are visits by a truck to a distribution centre or a courier delivering packages to a building. Currently these jobs are manually entered into the system. The objective of this project is to generate a list of jobs from vehicle point data, with a lower number of inaccuracies. That is few false positives and no false negatives. The rationale behind the project is to reduce the number of jobs which are entered manually. This could save customers of Telogis time and also allow customers that were not manually entering jobs to use the features that they enable. Although this project has been worked on by a team of students, each have specialised in a particular area. Matthew will be discussing data labelling and clustering of arbitrary point data. James will cover the machine learning methods used to form the prediction models. Dion will be demoing the reporting interface used to display the results of the project and change variables of the algorithms used.

Haydon Baddock

Wheelchair Data Logger and Analyser

The purpose of this project has been to determine the feasibility of accurately calculating the effective range of an electric wheelchair based on previously logged data from past usage of the chair. By logging battery usage statistics such as voltage and current at regular intervals, as well as location data, as a chair is being operated, it can be determined how much energy a not yet undertaken journey will consume by comparing the logged data with route and elevation data provided by the Google Maps API. The end result is an app that can be used to reasonably judge how much power a given route will consume, once enough usage has been recorded.

Jake Crouchley

Geological Viewer for iPhone – ARANZ Geo

For geologists around the world, ARANZ Geo Ltd’s “Leapfrog” branded modelling products are synonymous with high-end 3D visualisation. ARANZ has traditionally based its products on the desktop Windows platform, which otherwise dominates the mining sector. Feedback from existing customers has indicated a strong desire for to be able to access models for visualisation on mobile devices. The project was aimed at producing a proof-of-concept mobile application on the iOS platform that allowed users to open Leapfrog files and view them on an iPhone. Another aim was to experiment with new techniques for interacting with the 3D model through a touch interface.

Amy Martin

Forestry Apps for Deflection Planning and Tension Monitoring

I am developing two Android apps in conjunction with UC School of Forestry to support cable yarding operations in the forestry industry. The general objective of both apps is to address the issue of occupational safety by indicating the level of risk in the work environment. 1) The Deflection Planning app is essentially a calculator which trivialises a back-of-the envelope safety check to aid operational planning. When a forestry planner supplies physical dimensions of the cable yarding operation, the app computes an estimate of the deflection capacity. The deflection capacity is directly proportional to the safe loading limit and is a critical metric for configuration safety. 2) The Tension Monitor app is an interactive real-time plot of cable tension which completes an important feedback loop for the machine operator. A safe loading limit indicator and aggregated statistics give insight into loading cycles and extreme events to help improve efficiency and lower the cost of maintenance.

Daniel Tapp

Group Collaboration using Multi-Touch Tabletop Environments

Multi-touch displays are a new and quickly changing area of technology. Touch phones, tablets and computer screens are quickly becoming a standard feature in devices built for any environment and user. In recent years new hardware technology has made touch tabletop interactions a possibility, interactions that would have been limited to the magic of Hollywood a decade ago are now available on a multitude of easily accessible devices. This project introduces a platform for investigating how multi-touch tabletop displays can be used to enhance collaborative experiences for users in organizational and planning activities. An application has been developed so that a group of team members can simultaneously estimate stories or tasks in an agile planning poker format using a single tabletop display. The application takes the users through the pre-estimation, estimation, and negotiation stages of the estimation process.

Simon Crequer

Improved Driving App for Double Telepresence Robot

Double is a telepresence robot which lets users have a virtual presence in a remote place, such as an office or nursing home. The robot is controlled remotely via an app in a web browser. The app provides basic controls such as moving the Double using the arrow keys. Due to the latency between the driving app and the Double, it can be difficult to drive the Double. This project involves creating a new app for the Double and a new driving app, which allows for custom features to be added. New driving controls are added to improve the driving experience. The project also explores the idea of automatic navigation, to allow the Double to drive to preset locations.

Victor Chang

Interactives in the Computer Science Field Guide

Despite Computer Science fundamentals being introduced to the NCEA curriculum in 2011, many secondary level education institutes are reluctant to adopt these new achievement standards – partly due to a small proportion of secondary school level teachers feeling comfortable with teaching CS fundamentals. The Computer Science Field Guide is an open-source learning resource for both secondary school level students and teachers, containing educational information regarding all the concepts being taught in the NCEA standards. The CS Field Guide also contains “interactives” – browser based educational games that teach the user some aspect of CS concepts. These interactives were the focus of the project.The interactives that were developed over the course of the project related to human computer interaction, searching algorithms and text compression. They were developed with an effective learning experience in mind, applying education theory concepts such as the Zone of Proximal Development (ZPD), flow, constructivism, among others.

Su-Shing Chen

UAV Swarm: Collision Avoidance Outdoors

When a UAV is collaboratively flying through relatively confined spaces as part of a UAV swarm, there is a high risk of colliding with obstacles or other UAVs. This paper proposes a method to detect and avoid such collisions by predicting the movement of objects (static and dynamic) within the data provided by a depth-sensing camera. This research is a part of a team project enabling a swarm of UAVs to collaboratively map an environment with the future goal of navigating forests to prune trees.

Chris Marffy

UAV Traffic Management with Airways

New Zealand, as well as many other countries, is struggling with the rapid increase in usage of Unmanned Aerial Vehicles (UAVs). Airways New Zealand has developed a UAV traffic management (UTM) system in the form of their Airshare website (https://www.airshare.co.nz/). The system requests that people piloting UAVs log their flights so that airspace rules can be followed and air safety maintained. The purpose of the logging is to help Airways to manage the flights and to solve problems that may occur from UAVs flying when/where they should not be. This project covers the development of a mobile flight logging app for drone users, to complement the existing Airshare website with the convenience and extra features available from mobile devices.

Josh Norton

Robust Features for Accurate Pose Estimation (Trimble)

The goal of this project is to accurately detect the pose (<3mm) of an object 5m away using computer vision algorithms. The object is initially segmented from the background using an adaptive background subtraction approach. A Gaussian blur and a Canny edge detection algorithm are then applied to find the strongest edges. A probabilistic Hough line detection algorithm is used to find the optimal candidate horizontal lines from the set of edges. The most dominant horizontal line is weighted by proximity and length of candidate lines. The average vertical location of pixels falling on this line is then used as a key feature to continuously determine the vertical position and orientation of the object to within 3mm.

Monty Anderson

Design and Implementation of n-HARM and Visualisation

Security analysis in modern systems is an extremely tedious process. Traditionally it required the network administrators and security professionals to perform extensive analysis on a variety of large data sources and logs. Due to the tedious nature of this task and the difficulty in spotting trends through such an obscure medium, there has been considerable interest in network security modelling and visualisation. Put forward as a potential solution to this problem was a scalable, dynamic security analysis and visualisation system. This system would be built upon Safelite, a network security analysis framework. It would use Hierarchical Attack Representation Models (HARMs) as a scalable security model. The goal of this project is to design and implement n-layer HARMs, alongside the improvement of the system and its existing visualisation methods.

Michael Roman

User-Aware Web Based Tracking and Control

Many websites today have a large reliance on third party networks for the content being displayed on their web pages. With such a high inter-connectivity, users today have little indication on who has access to their browsing habits. This project aimed at creating a new browser add-on which both disable many of these connections as other add-ons do, but also familiarize with these possible privacy risks.

  • Phone: +64 3 364 2362
    Fax: +64 3 364 2569
    admin@cosc.canterbury.ac.nz
  • Computer Science and Software Engineering
    University of Canterbury
    Private Bag 4800, Christchurch
    New Zealand
  • Follow us
    FacebookYoutubetwitterLinked In