Skip to content

Latest commit

 

History

History
177 lines (151 loc) · 12.5 KB

README.md

File metadata and controls

177 lines (151 loc) · 12.5 KB

Overview

RIOT (REST IOT) is intended to simplify and componentize the code to build robotic kits with REST style web service. It is designed with flexibility and extensibility and is relatively easy to extend to general purpose IOT.

The followings are in the repository.

  • piServices package - an HTTP service with simple REST IOT protocol for client-server communication. It provides basic gpio, cpu, and memory data via HTTP. Extensibility example can be found in the picarServer.
  • iotServerLib package - a set of classes that encapsulates sensor/drive compoennts (for now mostly limited to what in Adeept Mars PiCar). Usage example can be found in the picarServer.
  • pySmartPlugService module - a RIOT service that provide remote access/control to TP-Link Kasa HS1xx smart plug.
  • videoStreaming package - an HTTP video streming service based on Raspberry Pi camera. Usage example can be found in the picarServer.
  • picarServer module - this is the server for Adeept Mars PiCar. Besides the otiginal TCP support it added HTTP support. It is (mostly) backward compatible with existing TCP client.
  • Riot class libs (.NET C#) - a set of class libraries that encapsulates sensor/device compoennts on the server via HTTP.
  • Simple Android app (Xamarin based) - to monitor and control Raspberry Pi, Kasa smart plug, Adeept Mars PiCar, and runs web service on the phone to serve sensor data and some actions
  • examples - contains simple example code for using the packages

Getting Started

  1. download/clone the respository. These steps assume that the code is under /home/pi/riot.
  2. Install the Python packages with pip3. This will install the Python packages required for PiCar Server.
cd /home/pi/riot
sudo python3 pip3 install .
  1. Configure and run PiCar server. Please refer the the Configuration section in picarServer/readme.md for details. Command to run picar server
python3 /home/pi/riot/picarServer/server.py /home/pi/riot/picarServer/picarconfig.txt
  1. Install and run smart plug RIOT service (see readme.md in the folder for more)
    1. install from https://github.com/vrachieru/tplink-smartplug-api. Alternatively, just download the code and copy the api.py to /home/pi/riot/pySmartPlugService and rename to tplink_smartplug.py.
    2. run smart plug service using one of these approaches
      • run as stand alone service - refer to readme.md under pySmartPlugService
      • run as part of piCarServer by adding the 2 lines to /home/pi/riot/picarServer/server.py as following
        picarHandler = PiCarRequestHandler(commandHandler, name='PiCar', authorizationList=authorizationList)
        kasaHandler = KasaHs1xxRequestHandler(name='PiServerSmartPlugs', authorizationList=authorizationList)
        kasaHandler.addSmartPlug('plug01', '192.168.1.85')  # replace 192.168.1.85 with the smart plug's IP address'
  1. Setup RiotDevices Android app
    1. Build and deploy to Android phone with VS
    2. Grant storage permission for the Devices.Android app
    3. Re-launch the Devices app
    4. Set the Pi Server IP address and password.
      1. goto "Pi Server". There would be an error message after timeout.
      2. click "Edit" then update the server Address (address:port) and Credential (user:password) fields
      3. click Save. This would create a devicessettings.txt file under docs folder.
      4. close & restart the app
    5. Setup the Pi Car server address and password following the same steps as above except by go to "Pi Car".
    6. Configure the PiCar settings through the "Settings"
      • Motor Speed: the speed of the motor (0 to 100)
      • Steering Angle: the angle for left/right steering servo (0 to 90)
      • Delta Camera Angle: the delta angle to increment when holding the head's right/left/top/down buttons
      • Config distance scan with start/end angle position and angle increment
    7. Config smart plug by following the same steps as 4.
    8. Alternatively, edit the devicessettings.txt file directly to change server IP, credential, and adding new servers new smart plugs. Start by copying the file from the RiotDevices folder to the Android device's docs folder.
    9. Change the commands by adding/updating the groups with Type = PiCommand.
    10. Configure phone web service by updateing the following settings
      • whether to enable the phone web service: EnableWebService=true
      • the port number for the web service: ServerPrefix=http://*:5678
      • the root path for service: ServiceRootPath=
      • the command root path for the service: ServiceActionRootPath=/cmd
      • list of user:password: ServiceCredentials= user1:password1, user2:password2

Details

RIOT Protocol

RIOT (REST for IOT) uses a simple HTTP REST protocol for client-server communication. For current version, the server always response with JSON.

HTTP Url

The url path is the unique identification to the target.

<scheme>://<host>/<root>/<device>/<node>/<data>[?parameters]
  • Scheme: http (will support https in the future)
  • Host: the server’s IP address or name and corresponding port specification
  • Root: the root path. In the Python RequestHandler, the root path is used to determine the instance of handler.
  • Device: the device ID or the path to the device if it is different than the root. In general, the // determines where the device is.
  • [Optional] node: the node ID or the path to the node separated with ‘/’ if it contains multiple layers.
  • [Optional] data: the data to get/post. the data is depended on the service.

HTTP Methods

  • Get – to retrieve information/data from the target server device. In general this does not alter the state of target server device.
  • Post and Put – to set value or send command to a target server device.

HTTP Body

The HTTP body is for PUT and POST only. It contains the value or content that is sent to the server target.

piServices - Python package

The Python piServices package implements the base classes to support RIOT HTTP protocol plus basic services for Raspberry Pi. It has dependency on Flask.

  • BaseRequestHandler - the base class for handling HTTP request. The class supports custom handler by implementing a handler derived from this base class. All instances of derived classes will be automatically registered (via init) with root paths (the device in the protocol) to handle.
    • doGet/doPut/doPost/doDelete - Derived classes implement these to handle corresponding request. By default, doPut calls doPost so they behave the same.
    • checkAuth - check authorization. Override to implement custom authorization.
    • processRequest - a class method that should be called to handle http requests. It finds the handler instances corresponding to the request's root path (device) from registered instances. The instance will be invoked to handle the requests for the root path (device).
  • PiGpioRequestHandler - this class is derived from BaseRequestHandler to handle requests for GET and POST from/to Raspberry Pi GPIO.
  • PiSysRequestHandler - this class is derived from BaseRequestHandler to handle requests for GET information from Raspberry Pi system (cpu, memory). The POST allows executing commands on the Pi.

pySmartPlugService - Python module

This is a RIOT service that provide remote access/control to TP-Link Kasa HS1xx smart plug. It allows clients send http requests to get status and control the plug's on/off and led. See readme.md in the folder for more.

iotServerLib - .NET class library

iotServerLib implements a set of class to encapsulate hardware components connected with Raspberry Pi's GPIO such as motor, servo, and others. It provides reusable building blocks and hide the details of the hardware. It is also intended to establish a standard interface that can encapsulate different hardware components/drivers with the same or similar functions.

  • IotNode - the base class for all IOT nodes (as in a composite pattern)
  • PiIotNode - the base class for all IOT nodes in Raspberry Pi
  • PiDcMotor - encapsulates a DC motor that connected via controller similar to L293D
  • PiServoI2c - encapsulates a servo connected via Adafruit_PCA9685
  • PiRGBLed - encapsulates an RGB LED connected through GPIO
  • PiUltrasonic - encapsulates an Ultrasonic sensor connected through GPIO
  • PiStripLed - encapsulates the RGB led strip using ws281x
  • PiLineTracking - encapsulates a line tracking sensor connected through GPIO
  • PiHygroThermoDHT - encapsulate the DHT sensor for humidity & temperature (dependency: Freenove_DHT.py)

videoStreaming - Python package

This package supports an HTTP video streming service based on Flask. The main code is from https://blog.miguelgrinberg.com/post/video-streaming-with-flask with addition of face tracking.

  • FaceTracker - detect and track faces with input image frames

videoStreaming works in either Windows and Pi:

  • Windows: just run python videoStreamingService.py
  • Pi: just run python3 videoStreamingService-pi.py

picarServer - - Python app

This is the server for Adeept Mars PiCar. By leveraging the above Python packages, it supports both HTTP RIOT protocol (based on Flask) and the original Adeept TCP implementation. It also added face tracking for the video streaming.

  • PiCarRequestHandler - derived from piServices.BaseRequestHandler to handle requests for the PiCar.
  • PiCar - the overall control for the Adeept Mars PiCar that contains child nodes:
    • PiCarDrive - controls the "drive train" for the picar - DC motor, drive direction servo, and both left/right RGB LEDs
    • PiCarHead - controls the "head" for the picar - ultrasonic sensor, camera, and both horizontal/vertical servos
    • line tracker using iotServerLib.PiLineTracking
    • strip LED using iotServerLib.PiStripLed
  • PiCarCamera - encapsulates the PI camera and video streaming with face tracking
  • PiCarCommandHandler - component to handle all the commands for both PiCarRequestHandler (HTTP) and picarTcpService (TCP)
  • picarTcpService - the original TCP server
  • Config - simple csv based configuration
  • server - the server module that loads configuration and starts all necessary threads for the picar

Rito and Riot.* - .NET class libraries for Riot

This is a set of C# libraries that can be used to build server and client app to control/communicate via RIOT HTTP protocol. It provides reusable building blocks and a standard interfaces that hides the details of the hardware.

Riot

This project defines all the base interfaces and classes for RIOT. Riot uses composite pattern to form the base interfaces and classes. Refer to other projects to use and extend the base classes.

Riot.IoDevice

The library implements RIOT clients for some IO devices/sensors such as DC motor, servo, DHT, and ultrasonic sensors.

Riot.Pi

The Riot.Pi wraps the Raspberry Pi's system and gpio in simple RIOT HTTP protocol. Riot.Pi contains the data and nodes for Raspberry Pi's system and gpio.

Riot.SmartPlug

The Riot.SmartPlug implements the client to access/control smart plugs through web service using RIOT HTTP protocol. Currently, only the Kasa HS1xx smart plugs are supported. It includes followings.

Riot.Phone

The Riot.Phone provides smartphone's sensor data via Riot web service. It also supports some actions that can be posted to the service to perform on the phone.

RiotDevices - Android App

The RiotDevices is a simple Android app to monitor and control devices via web service using RIOT HTTP protocol. The Android App is built with the Riot libraries and Xamarin Forms. It should be relative easy to make it work on iOS but had not tried on any IOS devices!!

  • Monitor the phone's sensors and device info
  • Start a web service on the phone to provide senor data remotely and some remote actions such as text to speech and vibrate
  • Monitor Raspberry Pi's CPU amd memory status.
  • Send system commands, like reboot and shutdown, can also posted to the Pi.
  • View Kasa smart plugs (model HS1xx) status and control on/off of the switch, reboot the plug, and set led-off status.
  • Control robotic PiCar (Adeept Mars PiCar)
  • Video streaming via HTTP.

Examples

  • picarServer has examples for the followings:
    • custom request handler class - PiCarRequestHandler
    • usage of iotServerLib - PiCar, PiCarDrive, PiCarHead
    • process HTTP requests with Flask - server.py
    • video streaming - picar.py, picarCamera.py
  • HTTP Request examples using HTTPie: httpieRequests.txt
  • Simple Raspberry Pi HTTP server: under piserver folder