cornellev/driverdash

Mobile driver dashboard for 2023 competition.

Attention! Are you a software or electrical team member at competition looking to troubleshoot the app? Are you just curious as to how to actually use it? Check out SETUP.md for instructions on how to get it up and running and what to do when things go wrong.

Driver Dash 2023

Jason Klein '26, Ari Kapelyan '25, Kalehiwot Dessalgne '25, Drew Wilenzick '26, Lauren Kam '25

A quick solution to a driver dashboard for competition in less than two weeks (!) The core components of the dashboard are (1) live statistics from the DAQ that allow it to communicate current speed and power consumption to the driver and (2) the ability to archive all sorts of data like GPS for later analysis. We also implemented location tracking and a labelled map of the track.

Upon receiving properly-formatted TCP requests, the phone will make corresponding updates to the stats on the dashboard and will save the requests and their timestamps onto the device. Saved requests are stored in a format looking like

{
  "2023-03-26 23:22:02.709000": {
    "rpm": 300,
    "voltage": 24.1,
    "safety": 1
  },
  "2023-03-26 23:22:02.711000": {
    "rpm": 300,
    "voltage": 24.1,
    "safety": 1
  }
  // ...
}

You can find the saved files by opening the Files app on the phone and navigating to On My iPhone > DriverDash. There you should find three files: daq.json, lord.json, and phone.json (stores GPS stuff from the phone). Note that, while the extension is .json, these aren't actually valid JSON files. the last comma needs to be replaced with a } in order for them to be valid JSON.

Setup

Since this is a native iOS app, it requires a Mac with XCode to build!

1. Make sure that you have Xcode development tools and cocoapods installed. If the initial instructions on the CocoaPods website don't work right away (like there's some issue with ActiveSupport or whatever), install a newer version with brew install ruby, and then attempt installation again using this version with sudo gem install -n /usr/local/bin cocoapods.
2. Navigate into the directory containing Podfile and run pod install from the terminal.
3. Run open driver-dash-2023.xcworkspace (not .xcodeproj)!
4. Connect the development phone, select it as a build target, and press build

Testing

The server aspect of the app can only be tested when the app is run from a physical phone; it will not work in the Apple Simulator app!

Assuming that the phone is connected to WiFi, you can find the phone's IP address by going to Settings > WiFi > (i) next to connected network name > IP Address, and the app will print to the console the port that the server will be running on (which should be 8080 because I hardcoded it). An IP address might look like 10.48.141.41. When connected over cellular, the IP address is most often 172.20.10.1. If that doesn't work, make sure that your laptop and the phone are connected to the same network and then run arp -a in a terminal. Pick the IP address with MAC address 72:81:eb:eb:aa:64.

Due to constraints imposed by the ESP32s, we have switched from HTTP/WebSockets to raw TCP requests. The server on the phone is now a simple TCP server listening for requests in a while true {} loop on a different thread. Test the TCP server by running python test.py 10.48.141.41 (replacing the actual IP address there). If all goes well, the dashboard should update with some new values, the connected indicators at the bottom of the phone UI should turn green for a bit, and there should be new content in the .json files.

test.py works by making a connection to the front and back daq servers and sending random data. Call it with --repetitions 5 to have it make five requests each time instead of one (works with any positive number).

Structure

DriverDash (the root view), DataView, and MapView control the appearance of the iPhone app, laying out the values and numbers and indicators and displaying the data. To handle state and effects, the project uses the Model-View-Controller (MVC) architecture, with DriverDash as the view.

DDModel is the model, and is responsible for storing the state for the app, such as the current speed and power values and the connection statuses. Every file within the Controllers/ directory is a controller responsible for managing distinct aspects of the app. DDLocation gets location data from the phone to use as a fallback, and DDServer runs the TCP servers on separate threads and updates the model when they receive information. Overall, the structure looks like this:

There are also utility scripts Coder and Serializer in Utilities/ which handle encoding/decoding JSON data and writing to files respectively. Because of the way that JSON parsing works in Swift, Coder will always have up-to-date typed structs FrontPacket, BackPacket, LordPacket outlining which data it expects to see sent to each of the three servers (it also has PhonePacket which is useful for encoding the phone's GPS data as JSON).

Competition Update

There were two major additions at competition: a lap counter, and a stopwatch.

Lap Counter was a simple button that increases the number displayed on the button by one. This number begins at 1 on each build. The use of this is because of the long length of the run, making it easy to lose count of what lap you're on (and getting this wrong would be detrimental to the energy efficiency of the run). This does require the driver to press a button, but they have plenty of time to do so (it's mostly just a tool for them!).

Stopwatch was a simple start/stop/reset stopwatch to count up from the beginning of the run. This is because we realized that we'd run out of time to connect the live-timing-dash to the app, which would allow people on the sidelines to press start and stop for the most accurate timing data. This is necessary though so that the driver knows how much time they have left from the 35 minutes allowed for a run. The code for this stopwatch was taken heavily from this site.