- Demo video
- What are we making
- Why are we making it / What problems does it solve
- How are we going to make it
- Features Overview
- Installation Guide(Set-up)
- User Manual
- Resources
- Development Team
We will endeavour to make a system to control the emulated car in the tool provided to us. We will write code for the emulated car to give it the capabilities we desire as well as to create an interface with which to control the car remotely & safely.
The following is a brief list of the Features of the system we plan to develop develop:
1: The car will be controllable using manually via interface & keyboard arrow keys.
2: The car will be controllable with smart voice commands.
3: The car will be controllable via a map with pathfinding functionalities.
By making the car be controllable with voice commands we hope to make it accessible to people unable to manually operate a computer and possibly more convenient. Autonomous navigation will make it much simpler for a user to move the car to a desired point effectively. They can do so with many fewer inputs than if they were to steer the car themselves. In adittion, currently with the panademic ravaging our world , it is sigificantly better to do more activities remotely with less risk of spreading viruses via human contacts or objects.
To build the system described above we are planning to use the following coding languages: Swift, Python and C++.
The reason we have chosen to use these different tools is to leverage the different strenghts of the utilized technologies as well as the individual strengths of our development team.
From a business standpoint we argue that these different interfaces are necessary for different users. Some drivers may want total control over the car, which is why we are enabling manual controls in our python application & arrow keys control. Others may want a more hands off experience with voice control, also in python. Lastly, some passengers may want a layed back and fully autonomous driving experience, which is why we are implementing the pathfinding app in swift.
We use Swift and SwiftUI to build the app which is responsible for presenting the map of the surrounding area as well as using pathfinding algorithms to find a path to the desired destination. Currently the app has two pathfinding algoriths, one that can quickly generate paths at the cost of them often being sub-optimal. And one that uses an a-star algorithm to generate the shortest path at the cost of it being slower to calculate. The app will then generate a set of instructions for the car to follow based on the calculated path. We may also implement a waypoint based system with predetermined routes between the different waypoints.
Python will be used to develop our alternative user interface, responsible for manual controls via the interface along with keyboard input, as well as smart voice recognition system to verbally deliver commands to the car and comprehend basic synonyms for driving purposes.
C++ together with MQTT will be the bridge between our various interfaces and the car(SMCE). Arduino code will be responsible for executing the several commands given to the car to perform accordingly.
There are a few things you need to set in place before kicking off:
- Download Arduino IDE
- Download SMCE-gd
- Download PyCharm
- Download Eclipse Mosquitto MQTT
- Clone the repository: https://github.com/DIT112-V21/group-08
- Install PyCharm Libraries
- Click on "File" Top-Left
- Select "Settings"
- Click on the arrow line of "group-08"
- Choose "Python Interpreter"
- Press on the "+" button to add the libraries (packages)
- Following instructions for MacOS only
- Install Xcode (also installs Swift and SwiftUI)
- If you do not already have an alternative for MQTT
- Install HomeBrew
- Install mosquitto by entering the following in the terminal: $ brew install mosquitto
- Open SMCE Remote.xcworkspace in Xcode and run the iPhone 12 Pro simulator.
It is important to use the simulator for this specific device since other screen sizes may have visual errors (which may or may not be fixed eventually). If this part works fine then ignore the following instructions.
- If any CocoaPods issues arise Install CocoaPods with following terminal commands
- $ sudo gem install cocoapods
- Navigate to xcode project directory
- $ pod install
If using an Apple Silicon computer try adding arch -x86_64 before each command if you are having issues.
- If any package dependency issues arise
- In Xcode navigate to File > Swift Packages > Add Package Dependency
- Select SMCE Remote as project
- Paste this URL: https://github.com/lucasbrown/swiftui-visual-effects
- Select master branch or latest release
As a driver using Python GUI :
-
You can manaully control the car movement from the buttons illustrated in the main window.
-
You can control the car speed from the "Vertical Slider" on the right, along with mointoring the current car speed from the bar.
-
You can control the car movement from your Keyboard Arrow Keys.
-
You can opt to Help Window from "How To Use" button, to see instructions for the app usage.
-
You can opt to Voice Control Window from"Voice Control" button , to control the car movement via voice commands.
-
You can see your previous & current history using the app from "Logging" Button.
-
You can instantly stop the car from "Brake" button.
-
You can control the car verbally by clicking on the button"Click,then say a command".
-
You can see instructions displayed to avoid any unclarities with pattern's usage.
Following instructions for swift pathfinder app
- First start the mosquitto server by entering the following in the terminal: /usr/local/sbin/mosquitto
- Then start the app in the iPhone 12 Pro simulator in Xcode
- Then start the arduino sketch in SMCE
- To begin, tap the center button on the home page of the app called "GRID" (or "BEGIN" depending on your version)
- To the bottom right of the screen there is a button with an A* icon. This will enable or disable the A-star algorithm used for the pathfinding. Disabling it will result is significantly faster path generation, but at the cost of these paths often being sub-optimal. If you are unhappy with the generated path or the app is unable to find a path at all, we suggest you enable the A-star algorithm.
- Tap on any of the white squares in the grid to generate a path, red squares represent unreachable terrain.
- After generating a path, the button to the bottom left is used to send the instructions generated to the car. Tapping this button will cause the car to move.
- You can repeat the two recent instructions how many times you want, but be aware that due to minor inaccuracies in the cars movement, the pathfinding will become less accurate over time. This may cause the car to get stuck or stop nowhere near its destination.
The following Software libraries, APIs and External services were used during the development of the PathFinder's Both Interfaces & General Functionalities:
- Arduino IDE
- SMCE-gd
- Smartcar Shield
- Xcode
- PyCharm
- PyCharm Libraries
- DialogFlow API
- Eclipse Mosquitto MQTT
- C++