Single Board Computer (SBC)
A powerful single board computer is a fundamental requirement to implement autonomy and create a collaborative network between various robots in a system. The processing unit needs to handle Ubuntu 14.04, ROS, Vision algorithms, and control an Arduino board, all running in parallel. To perform these tasks, it is important to choose a single board computer, which has capable enough hardware, in terms of processing power, RAM, etc., and supports other peripherals, such as a Kinect or Hokuyo Laser Scanner. For these reasons, performance, ease of integration, and support of peripherals have the highest weights in the trade study. Cost is an important factor because the system will require multiple units of these single board computers.
Single Board Computer Trade Study (Point Scale: 0-10)
| Weights | Raspberry Pi 2 | Edison | BeagleBone Black Rev C | Odroid XU-4 | |
|---|---|---|---|---|---|
| Performance | 25 | 6 | 10 | 4 | 10 |
| Cost | 10 | 10 | 4 | 7 | 6 |
| Documentation | 15 | 9 | 5 | 7 | 7 |
| Ease of Integration | 20 | 8 | 4 | 8 | 8 |
| Peripheral | 20 | 5 | 7 | 5 | 7 |
| Availability | 10 | 10 | 10 | 10 | 10 |
| Total | 100 | 7.5 | 6.9 | 6.4 | 8.2 |
Point-to-Point Communication
One of the core technologies that the project is intended to showcase is point-to-point communication between multiple vehicles. Additionally, the project will demonstrate how the vehicles can utilize shared data to collaborate in an effective manner. The network technology used to carry out this operation will have a huge impact on the implementation technique and overall performance of the system. Eliminating a central server is a non-negotiable requirement of the user, which is why cloud computing and wireless distributed computing are the two point-to-point communication options. Since this network will be supporting collaboration between real-time systems, low latency is a major requirement. Owing to the short timeline of the project, ease of implementation will play a critical role in getting the system up and running within the given timeframe.
Point-to-Point Communication Trade Study (Point Scale: 0-10)
| Weights | Cloud Computing | Wireless Distributed Computing | |
|---|---|---|---|
| Range | 15 | 10 | 7 |
| Reliability | 15 | 10 | 8 |
| Bandwidth | 5 | 10 | 8 |
| Latency | 25 | 6 | 9 |
| Cost | 5 | 8 | 8 |
| Ease of Implementation | 25 | 7 | 9 |
| Scalability | 10 | 8 | 6 |
| Total | 100 | 8 | 8.2 |
Communication Hardware
To implement the point-to-point communication system, each mobile platform needed a hardware to establish and join a network. The hardware that were considered were point-to-point WiFi routers like the Bullet 2HP and the XBee routers following the Zigbee or the Digimesh protocol. The key features desired in the communication units were – support for mesh networking, good indoor range, high bandwidth and robustness. All hardware that operated in the unlicensed frequency range were considered while selection.
Communication Hardware Trade Study (Point Scale: 0-10)
| Weights | XBee Pro 900 | XBee Pro | Bullet 2HP | |
|---|---|---|---|---|
| Low Frequency | 5 | 10 | 7 | 7 |
| Mesh Networking | 40 | 10 | 10 | 5 |
| Range | 20 | 7 | 3 | 5 |
| Reliability | 20 | 5 | 5 | 9 |
| Bandwidth | 5 | 5 | 5 | 9 |
| Form Factor | 10 | 7 | 7 | 3 |
| Total | 7.85 | 6.9 | 5.9 |
Mobile Platforms
The mobile platforms are the most critical component of the project. The mobile platforms are a central dependency for the project as every subsystem would be integrated with the platforms in terms of hardware and software. The team therefore considered the option of creating a platform from scratch to better suit the requirements or buy an off-the-shelf platform.
Mobile Platforms Trade Study (Point Scale: 0-10)
| Weight | Custom Platform | Oculus Prime | Pioneer | Husky | |
|---|---|---|---|---|---|
| Time To First Use | 15 | 5 | 8 | 9 | 6 |
| Cost | 25 | 6 | 8 | 7 | 1 |
| Non-Holonomic Steer | 10 | 9 | 5 | 5 | 5 |
| ROS Support | 20 | 1 | 9 | 7 | 9 |
| Support Documentation | 10 | 1 | 8 | 7 | 8 |
| Form Factor | 10 | 8 | 8 | 8 | 5 |
| Hardware | 10 | 7 | 8 | 5 | 7 |
| Total | 4.95 | 7.9 | 7.3 | 5.85 |