Difference between revisions of "Projects:2020s2-7352 Multi-Drone Tracking and Formation Control Platform"
| Line 1: | Line 1: | ||
== Introduction == | == Introduction == | ||
| − | + | Current development of UAV related systems are heavily dependent uponthe unique design of the intended UAV for which the system is developedfor. This leads to systems that are both limited to and limited by the buildof the UAV aswell as the environment it is flown in i.e. differences in sen-sory equipment, user interfaces and communication protocols. This has a negative impact on researchers work, narrowing its relevancy and applicability whilst encouraging research using simulated environments in order to navigate around limitations in physical environments.There are existing UAV products that aim to provide a common research platform however each are lacking in certain areas. BitCraze’s flagship drone,the CrazyFlie, is popular amongst research for its open ended control plat-form and support for both hardware and software extensions [1][2]. With this said, the CrazyFlie’s biggest downfall from a researchers perspective isits size. At 27 grams [3] the CrazyFlie limits users’ ability to attach additional sensory/communication equipment or larger batteries/motors without significant design changes. This is particularly relevant with new regulations1 | |
| + | that increase the weight limit of the micro RPA category from 100g to 250g,allowing researchers to fly heavier drones without the additional constraints applicable to larger categories [4]. The development of a heavier research purposed drone gives researchers the ability to use more powerful processors,sensors, communication modules, batteries and motors that open the door for studies using more robust and complex control algorithms.Further motivation for this project comes from the functional capabilities that are required to safely operate UAVs. Regardless of drone size or operator skill level, UAVs are capable of inflicting damage upon themselves and the environment around them. As such, effective control platforms are equipped with redundancies to avoid occurrences of collisions and mitigate potential injuries and costly equipment damages. What control platforms fail to consider is the variability in operator skill and flight environments.What this means is that relatively novice operators are able to fly as fast and as close to objects as experienced operators. The same goes for flights performed in a larger areas with no objects, versus smaller areas with many objects. Importantly, these shortcomings within current platforms are not impossible to overcome, however they subject researchers to additional time and costs that do not directly progress their research. By providing an adaptable control platform capable for wide ranging use with customization capabilities built in, we eliminate overheads, allow improvements/extensions to be reusable and provide a consistent base so that further research is relative and comparable | ||
== Project Team == | == Project Team == | ||
Latest revision as of 12:25, 7 October 2020
Contents
Introduction
Current development of UAV related systems are heavily dependent uponthe unique design of the intended UAV for which the system is developedfor. This leads to systems that are both limited to and limited by the buildof the UAV aswell as the environment it is flown in i.e. differences in sen-sory equipment, user interfaces and communication protocols. This has a negative impact on researchers work, narrowing its relevancy and applicability whilst encouraging research using simulated environments in order to navigate around limitations in physical environments.There are existing UAV products that aim to provide a common research platform however each are lacking in certain areas. BitCraze’s flagship drone,the CrazyFlie, is popular amongst research for its open ended control plat-form and support for both hardware and software extensions [1][2]. With this said, the CrazyFlie’s biggest downfall from a researchers perspective isits size. At 27 grams [3] the CrazyFlie limits users’ ability to attach additional sensory/communication equipment or larger batteries/motors without significant design changes. This is particularly relevant with new regulations1 that increase the weight limit of the micro RPA category from 100g to 250g,allowing researchers to fly heavier drones without the additional constraints applicable to larger categories [4]. The development of a heavier research purposed drone gives researchers the ability to use more powerful processors,sensors, communication modules, batteries and motors that open the door for studies using more robust and complex control algorithms.Further motivation for this project comes from the functional capabilities that are required to safely operate UAVs. Regardless of drone size or operator skill level, UAVs are capable of inflicting damage upon themselves and the environment around them. As such, effective control platforms are equipped with redundancies to avoid occurrences of collisions and mitigate potential injuries and costly equipment damages. What control platforms fail to consider is the variability in operator skill and flight environments.What this means is that relatively novice operators are able to fly as fast and as close to objects as experienced operators. The same goes for flights performed in a larger areas with no objects, versus smaller areas with many objects. Importantly, these shortcomings within current platforms are not impossible to overcome, however they subject researchers to additional time and costs that do not directly progress their research. By providing an adaptable control platform capable for wide ranging use with customization capabilities built in, we eliminate overheads, allow improvements/extensions to be reusable and provide a consistent base so that further research is relative and comparable
Project Team
Students
- Heath Rusby
- Alexander Materne
Supervisors
- Dr Hong Gunn Chew
Background
Control Platform
Custom Drone Prototype
Objectives
| Output | Description |
|---|---|
| Custom Built Drones | Drones will have a maximum weight of 250g, have a 30min flight time and have front and rear facing cameras. (minimum 2 drones) |
| Manual Drone Control | Drones will be able to be piloted by a person using a gaming controller. |
| Autonomous drone control | Drones will by capable of autonomous flight with python c++ and Matlab scripts including the ability to pass through stationary window on command. |
| Effective collision avoidance | Drones will be able to avoid virtual walls, other drones and static objects (During both Manual and Autonomous flight). The drone will be programmed with adjustable collision avoidance, such that an expert and a beginner will have different collision tolerances. |
| Thorough Documentation | All implementation and tests will be thoroughly documented as to allow continuity and usability for future users/developers |
Method
Results
Conclusion
References
[1] Business Insider. 2020. Drone Market Outlook: Industry Growth Trends, Market Stats And Forecast. [online] Available at: <https://www.businessinsider.com/drone-industry-analysis-market-trends-growth-forecasts?r=AU&IR=T>
