YoloV8 ONNX – Nvidia Jetson Orin Nano™ GPU CUDA Inferencing

The Seeedstudio reComputer J3011 has two processors an ARM64 CPU and an Nividia Jetson Orin 8G. To speed up inferencing with the Nividia Jetson Orin 8G with Compute Unified Device Architecture (CUDA) I built an Open Neural Network Exchange(ONNX) CUDA Execution Provider.

The Open Neural Network Exchange(ONNX) model used was trained on Roboflow Universe by Ugur ozdemir dataset which has 23696 images.

// load the app settings into configuration
var configuration = new ConfigurationBuilder()
      .AddJsonFile("appsettings.json", false, true)
.Build();

_applicationSettings = configuration.GetSection("ApplicationSettings").Get<Model.ApplicationSettings>();

Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss.fff} YoloV8 Model load: {_applicationSettings.ModelPath}");

YoloV8Builder builder = new YoloV8Builder();

builder.UseOnnxModel(_applicationSettings.ModelPath);

if (_applicationSettings.UseCuda)
{
   builder.UseCuda(_applicationSettings.DeviceId) ;
}

if (_applicationSettings.UseTensorrt)
{
   builder.UseTensorrt(_applicationSettings.DeviceId);
}

/*
builder.WithConfiguration(c =>
{
});
*/

/*
builder.WithSessionOptions(new Microsoft.ML.OnnxRuntime.SessionOptions()
{

});
*/

using (var image = await SixLabors.ImageSharp.Image.LoadAsync<Rgba32>(_applicationSettings.ImageInputPath))
using (var predictor = builder.Build())
{
   var result = await predictor.DetectAsync(image);

   Console.WriteLine();
   Console.WriteLine($"Speed: {result.Speed}");
   Console.WriteLine();

   foreach (var prediction in result.Boxes)
   {
      Console.WriteLine($" Class {prediction.Class} {(prediction.Confidence * 100.0):f1}% X:{prediction.Bounds.X} Y:{prediction.Bounds.Y} Width:{prediction.Bounds.Width} Height:{prediction.Bounds.Height}");
   }

   Console.WriteLine();

   Console.WriteLine($" {DateTime.UtcNow:yy-MM-dd HH:mm:ss.fff} Plot and save : {_applicationSettings.ImageOutputPath}");

   using (var imageOutput = await result.PlotImageAsync(image))
   {
      await imageOutput.SaveAsJpegAsync(_applicationSettings.ImageOutputPath);
   }
}

When configured to run the YoloV8.Coprocessor.Detect.Image on the ARM64 CPU the average inference time was 729 mSec.

The first time ran the YoloV8.Coprocessor.Detect.Image application configured to use CUDA for inferencing it failed badly.

The YoloV8.Coprocessor.Detect.Image application was then configured to use CUDA and the average inferencing time was 85mSec.

It took a couple of weeks to get the YoloV8.Coprocessor.Detect.Image application inferencing on the Nividia Jetson Orin 8G coprocessor and this will be covered in detail in another posts.

YoloV8 ONNX – Nvidia Jetson Orin Nano™ ARM64 CPU Inferencing

I configured the demonstration Ultralytics YoloV8 object detection(yolov8s.onnx) console application to process a 1920×1080 image from a security camera on my desktop development box (13th Gen Intel(R) Core(TM) i7-13700 2.10 GHz with 32.0 GB)

Object Detection sample application running on my development box

A Seeedstudio reComputer J3011 uses a Nividia Jetson Orin 8G and looked like a cost-effective platform to explore how a dedicated Artificial Intelligence (AI) co-processor could reduce inferencing times.

To establish a “baseline” I “published” the demonstration application on my development box which created a folder with all the files required to run the application on the Seeedstudio reComputer J3011 ARM64 CPU. I had to manually merge the “User Secrets” and appsettings.json files so the camera connection configuration was correct.

The runtimes folder contained a number of folders with the native runtime files for the supported Open Neural Network Exchange(ONNX) platforms

Object Detection application publish runtimes folder

This Nividia Jetson Orin ARM64 CPU requires the linux-arm64 ONNX runtime which was “automagically” detected. (in previous versions of ML.Net the native runtime had to be copied to the execution directory)

Linux ONNX ARM64 runtime

The final step was to use the demonstration Ultralytics YoloV8 object detection(yolov8s.onnx) console application to process a 1920×1080 image from a security camera on the reComputer J3011 (6-core Arm® Cortex®64-bit CPU 1.5Ghz processor)

Object Detection sample application running on my Seeedstudio reComputer J3011

When I averaged the pre-processing, inferencing and post-processing times for both devices over 20 executions my development box was much faster which was not a surprise. Though the reComputer J3011 post processing times were a bit faster than I was expecting

ARM64 CPU Preprocess 0.05s Inference 0.31s Postprocess 0.05

.NET nanoFramework Seeedstudio HM3301 library on Github

The source code of my .NET nanoFramework Seeedstudio Grove – Laser PM2.5 Dust Sensor HM3301 library is now available on GitHub. I have tested the library and sample application with Sparkfun Thing Plus and ST Micro STM32F7691 Discovery devices. (I can validate on more platform configurations if there is interest).

Important: make sure you setup the I2C pins especially on ESP32 Devices before creating the I2cDevice,

SHT20 +STM32F769 Discovery test rig

The .NET nanoFramework device libraries use a TryGet… pattern to retrieve sensor values, this library throws an exception if reading a sensor value fails. I’m not certain which approach is “better” as reading the Seeedstudio Grove – Laser PM2.5 Dust Sensor has never failed. The only time reading the “values” buffer failed was when I unplugged the device which I think is “exceptional”.

//---------------------------------------------------------------------------------
// Copyright (c) April 2023, devMobile Software
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// nanoff --target ST_STM32F769I_DISCOVERY --update 
// nanoff --platform ESP32 --serialport COM7 --update
//
//---------------------------------------------------------------------------------
#define ST_STM32F769I_DISCOVERY 
//#define  SPARKFUN_ESP32_THING_PLUS
namespace devMobile.IoT.Device.SeeedstudioHM3301
{
    using System;
    using System.Device.I2c;
    using System.Threading;

#if SPARKFUN_ESP32_THING_PLUS
    using nanoFramework.Hardware.Esp32;
#endif

    class Program
    {
        static void Main(string[] args)
        {
            const int busId = 1;

            Thread.Sleep(5000);

#if SPARKFUN_ESP32_THING_PLUS
            Configuration.SetPinFunction(Gpio.IO23, DeviceFunction.I2C1_DATA);
            Configuration.SetPinFunction(Gpio.IO22, DeviceFunction.I2C1_CLOCK);
#endif
            I2cConnectionSettings i2cConnectionSettings = new(busId, SeeedstudioHM3301.DefaultI2cAddress);

            using I2cDevice i2cDevice = I2cDevice.Create(i2cConnectionSettings);
            {
                using (SeeedstudioHM3301 seeedstudioHM3301 = new SeeedstudioHM3301(i2cDevice))
                {
                    while (true)
                    {
                        SeeedstudioHM3301.ParticulateMeasurements particulateMeasurements = seeedstudioHM3301.Read();

                        Console.WriteLine($"Standard PM1.0: {particulateMeasurements.Standard.PM1_0} ug/m3   PM2.5: {particulateMeasurements.Standard.PM2_5} ug/m3  PM10.0: {particulateMeasurements.Standard.PM10_0} ug/m3 ");
                        Console.WriteLine($"Atmospheric PM1.0: {particulateMeasurements.Atmospheric.PM1_0} ug/m3   PM2.5: {particulateMeasurements.Atmospheric.PM2_5} ug/m3  PM10.0: {particulateMeasurements.Standard.PM10_0} ug/m3");

                        // Always 0, checked payload so not a conversion issue. will check in Seeedstudio forums
                        // Console.WriteLine($"Count 0.3um: {particulateMeasurements.Count.Diameter0_3}/l 0.5um: {particulateMeasurements.Count.Diameter0_5} /l 1.0um : {particulateMeasurements.Count.Diameter1_0}/l 2.5um : {particulateMeasurements.Count.Diameter2_5}/l 5.0um : {particulateMeasurements.Count.Diameter5_0}/l 10.0um : {particulateMeasurements.Count.Diameter10_0}/l");

                        Thread.Sleep(new TimeSpan(0,1,0));
                    }
                }
            }
        }
    }
}

I’m going to soak test the library for a week to check that is working okay, then most probably refactor the code so it can be added to the nanoFramework IoT.Device Library repository.

.NET nanoFramework Seeedstudio HM3301 Basic connectivity

This is a “throw away” .NET nanoFramework application for investigating how Seeedstudio Grove HM3301 Inter Integrated Circuit bus(I²C) connectivity works.

Seeedstudio Grove HM3301 Sensor

My test setup is a simple .NET nanoFramework console application running on an STM32F7691 Discovery board.

Seeedstudio Grove HM3301 + STM32F769 Discovery test rig

The HM3301I2C application has lots of magic numbers from the HM3301 datasheet and is just a tool for exploring how the sensor works.

public static void Main()
{
    I2cConnectionSettings i2cConnectionSettings = new(1, 0x40);

    // i2cDevice.Dispose
    I2cDevice i2cDevice = I2cDevice.Create(i2cConnectionSettings);

    while (true)
    {
        byte[] writeBuffer = new byte[1];
        byte[] readBuffer = new byte[29];

        writeBuffer[0] = 0x88;

        i2cDevice.WriteRead(writeBuffer, readBuffer);

        //i2cDevice.WriteByte(0x88);
        //i2cDevice.Read(readBuffer);

        ushort standardParticulatePm1 = (ushort)(readBuffer[4] << 8);
        standardParticulatePm1 |= readBuffer[5];

        ushort standardParticulatePm25 = (ushort)(readBuffer[6] << 8);
        standardParticulatePm25 |= readBuffer[7];

        ushort standardParticulatePm10 = (ushort)(readBuffer[8] << 8);
                standardParticulatePm10 |= readBuffer[9];

        Console.WriteLine($"{DateTime.UtcNow:HH:mm:ss} Standard particulate    PM 1.0: {standardParticulatePm1}  PM 2.5: {standardParticulatePm25}  PM 10.0: {standardParticulatePm10} ug/m3");

        ushort atmosphericPm1 = (ushort)(readBuffer[10] << 8);
        atmosphericPm1 |= readBuffer[11];

        ushort atmosphericPm25 = (ushort)(readBuffer[12] << 8);
        atmosphericPm25 |= readBuffer[13];

        ushort atmosphericPm10 = (ushort)(readBuffer[14] << 8);
        atmosphericPm10 |= readBuffer[15];

        Console.WriteLine($"{DateTime.UtcNow:HH:mm:ss} Atmospheric particulate PM 1.0: {atmosphericPm1:3}  PM 2.5: {atmosphericPm25}  PM 10.0: {atmosphericPm10} ug/m3");


        ushort particulateCountPm03 = (ushort)(readBuffer[16] << 8);
        particulateCountPm03 |= readBuffer[17];

        ushort particulateCountPm05 = (ushort)(readBuffer[18] << 8);
        particulateCountPm05 |= readBuffer[19];

        ushort particulateCountPm1 = (ushort)(readBuffer[20] << 8);
        particulateCountPm1 |= readBuffer[21];

        Console.WriteLine($"{DateTime.UtcNow:HH:mm:ss} Particulate count       PM 0.3: {particulateCountPm03:3}  PM 0.5: {particulateCountPm05}  PM 1.0: {particulateCountPm1} ug/m3");


        ushort particleCountPm25 = (ushort)(readBuffer[22] << 8);
        particleCountPm25 |= readBuffer[23];

        ushort particleCountPm5 = (ushort)(readBuffer[24] << 8);
        particleCountPm5 |= readBuffer[25];

        ushort particleCountPm10 = (ushort)(readBuffer[26] << 8);
        particleCountPm10 |= readBuffer[27];

        Console.WriteLine($"{DateTime.UtcNow:HH:mm:ss} Particle count/0.1L     PM 2.5: {particleCountPm25}  PM 5.0: {particleCountPm5}  PM 10.0: {particleCountPm10} particles/0.1L");


        byte checksum = 0;
        for (int i = 0; i < readBuffer.Length - 1; i++)
        {
            checksum += readBuffer[i];
        }
        Console.WriteLine($"{DateTime.UtcNow:HH:mm:ss} Checksum payload:{readBuffer[28]} calculated:{checksum}");
        Console.WriteLine("");

        Thread.Sleep(5000);
    }
}

The unpacking of the value standard particulate, particulate count and particle count values is fairly repetitive, but I will fix it in the next version.

Visual Studio 2022 Debug output

The checksum calculation isn’t great even a simple cyclic redundancy check(CRC) would be an improvement on summing the 28 bytes of the payload.

.NET nanoFramework Qorvo DW1000 – Makerfabs Device SPI

The Makerfabs ESP32 UWB(Ultra Wideband) module has a Qorvo DW1000 and Espressif ESP32 module. The Espressif ESP32 module can run the .NET nanoFramework but does not have a Qorvo DW1000 library. (March2023)

Makerfabs ESP32 UWB(Ultra Wide Band) module

Before any coding I used nanoff to “flash” the Espressif ESP32 module with the latest version of .NET nanoFramework

Flashing Makerfabs ESP32 UWB module with nanoff

The Qorvo DW1000 module has a Serial Peripheral Interface (SPI) so the Master In Slave Out(MISO), Master Out Slave In(MOSI), Serial Clock(SCLK) and Chip Slave Select(CSS) pins have to be configured using the Configuration.SetPinFunction method of the nanoFramework.Hadware.Esp32 library

Makerfabs ESP32 UWB module schematic

Even though SPI is an industry standard there are often subtle differences which need to be taken into account when reading from/writing to registers. The DW1000 has a static “Device Identifier” which I used to debug my “proof of concept” code.

DW1000 Datasheet Register Map documentation for Register 0x00

The DeviceSPI program reads register 0x00 and then displays the decoded payload.

public class Program
{
#if MAKERFABS_ESP32UWB
    private const int SpiBusId = 1;
    private const int chipSelectLine = Gpio.IO04;
#endif

    public static void Main()
    {
        Thread.Sleep(5000);

        Debug.WriteLine("devMobile.IoT.Dw1000.ShieldSPI starting");

        try
        {
#if MAKERFABS_ESP32UWB
            Configuration.SetPinFunction(Gpio.IO19, DeviceFunction.SPI1_MISO);
            Configuration.SetPinFunction(Gpio.IO23, DeviceFunction.SPI1_MOSI);
            Configuration.SetPinFunction(Gpio.IO18, DeviceFunction.SPI1_CLOCK);
#endif
            var settings = new SpiConnectionSettings(SpiBusId, chipSelectLine)
            {
                ClockFrequency = 2000000,
                Mode = SpiMode.Mode0,
            };

            using (SpiDevice device = SpiDevice.Create(settings))
            {
                Thread.Sleep(500);

                while (true)
                {
                    /*
                    byte[] writeBuffer = new byte[] { 0x0, 0x0, 0x0, 0x0, 0x0 }; // 0x0 = DEV_ID
                    byte[] readBuffer = new byte[writeBuffer.Length];

                    device.TransferFullDuplex(writeBuffer, readBuffer); // 15, 48, 1, 202, 222
                    */

                    byte[] writeBuffer = new byte[] { 0x0 }; // 0x0 = DEV_ID
                    byte[] readBuffer = new byte[5];

                    device.TransferFullDuplex(writeBuffer, readBuffer); // 15, 48, 1, 202, 222
                       
                    uint ridTag = (uint)(readBuffer[4]<< 8 | readBuffer[3]);
                    byte model = readBuffer[2];
                    byte ver = (byte)(readBuffer[1] >> 4);
                    byte rev = (byte)(readBuffer[1] & 0x0f);

                    Debug.WriteLine(String.Format($"RIDTAG 0x{ridTag:X2} MODEL 0x{model:X2} VER 0X{ver:X2} REV 0x{rev:X2}"));

                    Thread.Sleep(10000);
                }
            }
        }
        catch (Exception ex)
        {
            Debug.WriteLine(ex.Message);
        }
    }
}
Visual Studio 2022 Debug window displaying the decoded value from Register 0x0

The DW1000 User Manual is > 240 pages, with roughly 140 pages of detailed documentation about the DW1000 register set so progress will be slow.

.NET nanoFramework RAK11200 – I2C SHT3C & SHT31

The RAKwireless RAK11200 WisBlock WiFi Module module is based on an Expressif ESP32 processor which is supported by the .NET nanoFramework and I wanted to explore the different ways Inter-Integrated Circuit(I2C) devices could be connected.

The RAK11200 WisBlock WiFi Module has two I2C ports and on the RAK5005 WisBlock Base Board the Wisblock Sensor, and RAK1920 WisBlock Sensor Adapter Module Grove Socket are connected to I2C1.

RAK11200 Schematic

The I2C1 the SDA(serial data) and SCL(serial clock line) have to be mapped to physical pins on the RAK11200 WisBlock WiFi Module using the nanoFramework ESP32 support NuGet. package

                Configuration.SetPinFunction(Gpio.IO04, DeviceFunction.I2C1_DATA);
                Configuration.SetPinFunction(Gpio.IO05, DeviceFunction.I2C1_CLOCK)

The first sample project uses a RAK1901 SHTC3 WisBlock Sensor because it plugs into the RAK5005 WisBlock Base Board.

RAK5005 Baseboard, RAK1901 Sensor and RAK11200 Core WisBlock modules
public static void Main()
{
    Debug.WriteLine("devMobile.IoT.RAK.Wisblock.SHTC3 starting");

    try
    {
        Configuration.SetPinFunction(Gpio.IO04, DeviceFunction.I2C1_DATA);
        Configuration.SetPinFunction(Gpio.IO05, DeviceFunction.I2C1_CLOCK);

        I2cConnectionSettings settings = new(1, Shtc3.DefaultI2cAddress);

        using (I2cDevice device = I2cDevice.Create(settings))
        using (Shtc3 shtc3 = new(device))
        {
            while (true)
            {
                if (shtc3.TryGetTemperatureAndHumidity(out var temperature, out var relativeHumidity))
                {
                    Debug.WriteLine($"Temperature {temperature.DegreesCelsius:F1}°C  Humidity {relativeHumidity.Value:F0}%");
                }

                Thread.Sleep(10000);
            }
        }
    }
    catch (Exception ex)
    {
        Debug.WriteLine($"SHTC3 initialisation or read failed {ex.Message}");

        Thread.Sleep(Timeout.Infinite);
    }
}
Visual Studio Output window displaying SHT3C temperature & humidity values

The second sample uses a Seeedstudio Grove – Temperature & Humidity Sensor (SHT31) pluged into a RAK1920 Sensor Adapter for Click, QWIIC and Grove Modules.

RAK5005 Baseboard, RAK1920 Sensor, RAK11200 Core WisBlock modules and Seeedstudio Grove SHT31
public static void Main()
{
    Debug.WriteLine("devMobile.IoT.RAK.Wisblock.SHT31 starting");

    try
    {
        Configuration.SetPinFunction(Gpio.IO04, DeviceFunction.I2C1_DATA);
        Configuration.SetPinFunction(Gpio.IO05, DeviceFunction.I2C1_CLOCK);

        I2cConnectionSettings settings = new(1, (byte)I2cAddress.AddrLow);

        using (I2cDevice device = I2cDevice.Create(settings))
        using (Sht3x sht31 = new(device))
        {

            while (true)
            {
                var temperature = sht31.Temperature;
                var relativeHumidity = sht31.Humidity;

                Debug.WriteLine($"Temperature {temperature.DegreesCelsius:F1}°C  Humidity {relativeHumidity.Value:F0}%");

                Thread.Sleep(10000);
            }
        }
    }
    catch (Exception ex)
    {
        Debug.WriteLine($"SHT31 initialisation or read failed {ex.Message}");

        Thread.Sleep(Timeout.Infinite);
     }
}
Visual Studio Output window displaying SHT31 temperature & humidity values

The SHTC3 and SHT31 sensors were used because they both have nanoFramework.IoTDevice library support.

Grove Base Hat for Raspberry PI with .NET Core 5.0

Over the weekend I ported my Windows 10 IoT Core library for Seeedstudio Grove Base Hat for RPI Zero and Grove Base Hat for Raspberry Pi to .NET Core 5.

RaspberryP and RaspberryPI Zero testrig

I have included sample application to show how to use the library

namespace devMobile.IoT.NetCore.GroveBaseHat
{
	using System;
	using System.Device.I2c;
	using System.Threading;

	class Program
	{
		static void Main(string[] args)
		{
			// bus id on the raspberry pi 3
			const int busId = 1;

			I2cConnectionSettings i2cConnectionSettings = new(busId, AnalogPorts.DefaultI2cAddress);

			using (I2cDevice i2cDevice = I2cDevice.Create(i2cConnectionSettings))
			using (AnalogPorts AnalogPorts = new AnalogPorts(i2cDevice))
			{
				Console.WriteLine($"{DateTime.Now:HH:mm:SS} Version:{AnalogPorts.Version()}");
				Console.WriteLine();

				double powerSupplyVoltage = AnalogPorts.PowerSupplyVoltage();
				Console.WriteLine($"{DateTime.Now:HH:mm:SS} Power Supply Voltage:{powerSupplyVoltage:F2}v");

				while (true)
				{
					double value = AnalogPorts.Read(AnalogPorts.AnalogPort.A0);
					double rawValue = AnalogPorts.ReadRaw(AnalogPorts.AnalogPort.A0);
					double voltageValue = AnalogPorts.ReadVoltage(AnalogPorts.AnalogPort.A0);

					Console.WriteLine($"{DateTime.Now:HH:mm:SS} Value:{value:F2} Raw:{rawValue:F2} Voltage:{voltageValue:F2}v");
					Console.WriteLine();

					Thread.Sleep(1000);
				}
			}
		}
	}
}

The GROVE_BASE_HAT_RPI and GROVE_BASE_HAT_RPI_ZERO are used to specify the number of available analog ports.

RAK3172LoRaWAN-NetCore on Github

The source code for a late beta version of my .Net Core C# library for RAK3172 modules is now available on GitHub.

Using SerialDataReceivedEventHandler was causing memory management problems so the core command processor now runs in its own Thread.(Though I worry about the continuous System.TimeoutExceptions)

A sample application which shows how to connect using Over the Air Activation(OTAA) or Activation By Personalisation(ABP) then send and receive byte array/hexadecimal messages.

//---------------------------------------------------------------------------------
// Copyright (c) September 2021, devMobile Software
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// Must have one of following options defined in the project\build definitions
//    PAYLOAD_BCD or PAYLOAD_BYTES
//    OTAA or ABP
//
// Optional definitions
//    CONFIRMED For confirmed messages
//		DEVEUI_SET
//
//---------------------------------------------------------------------------------
namespace devMobile.IoT.LoRaWAN.NetCore.RAK3172
{
	using System;
	using System.IO.Ports;
	using System.Threading;


	public class Program
	{
		private const string SerialPortId = "/dev/ttyS0";
		private const LoRaClass Class = LoRaClass.A;
		private const string Band = "8-1";
		private const byte MessagePort = 10;
		private static readonly TimeSpan MessageSendTimerDue = new TimeSpan(0, 0, 15);
		private static readonly TimeSpan MessageSendTimerPeriod = new TimeSpan(0, 5, 0);
		private static Timer MessageSendTimer ;
		private const int JoinRetryAttempts = 2;
		private const int JoinRetryIntervalSeconds = 10;
#if PAYLOAD_BCD
		private const string PayloadBcd = "48656c6c6f204c6f526157414e"; // Hello LoRaWAN in BCD
#endif
#if PAYLOAD_BYTES
		private static readonly byte[] PayloadBytes = { 0x48, 0x65 , 0x6c, 0x6c, 0x6f, 0x20, 0x4c, 0x6f, 0x52, 0x61, 0x57, 0x41, 0x4e}; // Hello LoRaWAN in bytes
#endif

		public static void Main()
		{
			Result result;

			Console.WriteLine("devMobile.IoT.LoRaWAN.NetCore.RAK3172 RAK3712LoRaWANDeviceClient starting");

			Console.WriteLine($"Serial ports:{String.Join(",", SerialPort.GetPortNames())}");

			try
			{
				using (Rak3172LoRaWanDevice device = new Rak3172LoRaWanDevice())
				{
					result = device.Initialise(SerialPortId, 9600, Parity.None, 8, StopBits.One);
					if (result != Result.Success)
					{
						Console.WriteLine($"Initialise failed {result}");
						return;
					}

					MessageSendTimer = new Timer(SendMessageTimerCallback, device,Timeout.Infinite, Timeout.Infinite);

					device.OnJoinCompletion += OnJoinCompletionHandler;
					device.OnReceiveMessage += OnReceiveMessageHandler;
#if CONFIRMED
					device.OnMessageConfirmation += OnMessageConfirmationHandler;
#endif

#if DEVEUI_SET
					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} DevEUI {Config.devEui}");
					result = device.DeviceEui(Config.devEui);
					if (result != Result.Success)
					{
						Console.WriteLine($"DevEUI failed {result}");
						return;
					}
#endif

					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Class {Class}");
					result = device.Class(Class);
					if (result != Result.Success)
					{
						Console.WriteLine($"Class failed {result}");
						return;
					}

					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Band {Band}");
					result = device.Band(Band);
					if (result != Result.Success)
					{
						Console.WriteLine($"Region failed {result}");
						return;
					}

					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} ADR On");
					result = device.AdrOn();
					if (result != Result.Success)
					{
						Console.WriteLine($"ADR on failed {result}");
						return;
					}

#if CONFIRMED
               Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Confirmed");
               result = device.UplinkMessageConfirmationOn();
               if (result != Result.Success)
               {
                  Console.WriteLine($"Confirm on failed {result}");
                  return;
               }
#else
					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Unconfirmed");
					result = device.UplinkMessageConfirmationOff();
					if (result != Result.Success)
					{
						Console.WriteLine($"Confirm off failed {result}");
						return;
					}
#endif

#if OTAA
					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} OTAA");
					result = device.OtaaInitialise(Config.JoinEui, Config.AppKey);
					if (result != Result.Success)
					{
						Console.WriteLine($"OTAA Initialise failed {result}");
						return;
					}
#endif

#if ABP
               Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} ABP");
               result = device.AbpInitialise(Config.DevAddress, Config.NwksKey, Config.AppsKey);
               if (result != Result.Success)
               {
                  Console.WriteLine($"ABP Initialise failed {result}");
                  return;
               }
#endif

					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Join start");
					result = device.Join(JoinRetryAttempts, JoinRetryIntervalSeconds);
					if (result != Result.Success)
					{
						Console.WriteLine($"Join failed {result}");
						return;
					}
					Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Join started");

					Thread.Sleep(Timeout.Infinite);
				}
			}
			catch (Exception ex)
			{
				Console.WriteLine(ex.Message);
			}
		}

		private static void OnJoinCompletionHandler(bool result)
		{
			Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Join finished:{result}");

			if (result)
			{ 
				MessageSendTimer.Change(MessageSendTimerDue, MessageSendTimerPeriod);
			}
		}

		private static void SendMessageTimerCallback(object state)
		{
			Rak3172LoRaWanDevice device = (Rak3172LoRaWanDevice)state;

#if PAYLOAD_BCD
			Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} port:{MessagePort} payload BCD:{PayloadBcd}");
			Result result = device.Send(MessagePort, PayloadBcd );
#endif
#if PAYLOAD_BYTES
			Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} port:{MessagePort} payload bytes:{Rak3172LoRaWanDevice.BytesToBcd(PayloadBytes)}");
         Result result = device.Send(MessagePort, PayloadBytes);
#endif
			if (result != Result.Success)
			{
				Console.WriteLine($"Send failed {result}");
			}
		}

#if CONFIRMED
		private static void OnMessageConfirmationHandler()
      {
			Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Send successful");
		}
#endif

		private static void OnReceiveMessageHandler(byte port, int rssi, int snr, string payload)
		{
			byte[] payloadBytes = Rak3172LoRaWanDevice.HexToByes(payload); // Done this way so both conversion methods tested

			Console.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Receive Message RSSI:{rssi} SNR:{snr} Port:{port} Payload:{payload} PayLoadBytes:{BitConverter.ToString(payloadBytes)}");
		}
	}
}

I have added XML Documentation comments which will need some rework, after I have “soak tested” the code for at least a week.

I have also added a method so the DevEUI can be set (intended for use after device firmware has been updated), fixed up my mistake with Binary Coded Decimal(BCD) vs. Hexadecimal strings.

I will also go back and apply the “learnings” from this refactoring to my other LoRaWAN module and platform libraries

.NET Core RAK3172 LoRaWAN library Part1

Basic connectivity

Over the weekend I have been working on a .NET Core C# library for the RAKwireless RAK3172 module using a RAK3172S breakout board, Seeedstudio Grove Base Hat for Raspberry PI and a Seeedstudio Grove-4 pin Female Jumper to Grove 4 pin Conversion Cable.

Raspberry Pi3 with Grove Base Hat and RAK3172 Breakout (using UART2)

The RaspberryPI OS is a bit more strict than the other devices I use about port access. To allow my .Net Core application to access a serial port I connected to the device with ExtraPutty, then ran the RaspberyPI configuration tool, from the command prompt with “sudo raspi-config”

RaspberyPI OS Software Configuration tool mains screen
RaspberryPI OS IO Serial Port configuration
Raspberry PI OS disabling remote serial login shell
RaspberryPI OS enabling serial port access

Once serial port access was enabled I could enumerate them with SerialPort.GetPortNames() which is in the System.IO.Ports NuGet package. My sample code has compile time options for synchronous and asynchronous operation.

//---------------------------------------------------------------------------------
// Copyright (c) September 2021, devMobile Software
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
//---------------------------------------------------------------------------------
namespace devMobile.IoT.NetCore.RAK3172.ShieldSerial
{
	using System;
	using System.Diagnostics;
	using System.IO.Ports;
	using System.Threading;

	public class Program
	{
		private const string SerialPortId = "/dev/ttyS0";

		public static void Main()
		{
			SerialPort serialPort;

			Debug.WriteLine("devMobile.IoT.NetCore.Rak3172.pHatSerial starting");

			Debug.WriteLine(String.Join(",", SerialPort.GetPortNames()));

			try
			{
				serialPort = new SerialPort(SerialPortId);

				// set parameters
				serialPort.BaudRate = 9600;
				serialPort.Parity = Parity.None;
				serialPort.DataBits = 8;
				serialPort.StopBits = StopBits.One;
				serialPort.Handshake = Handshake.None;

				serialPort.ReadTimeout = 1000;

				serialPort.NewLine = "\r\n";

				serialPort.Open();

#if SERIAL_ASYNC_READ
				serialPort.DataReceived += SerialDevice_DataReceived;
#endif

				while (true)
				{
					serialPort.WriteLine("AT+VER=?");

#if SERIAL_SYNC_READ
					// Read the response
					string response = serialPort.ReadLine();
					Debug.WriteLine($"RX:{response.Trim()} bytes:{response.Length}");

					// Read the blank line
					response = serialPort.ReadLine();
					Debug.WriteLine($"RX:{response.Trim()} bytes:{response.Length}");

					// Read the result
					response = serialPort.ReadLine();
					Debug.WriteLine($"RX:{response.Trim()} bytes:{response.Length}");
#endif

					Thread.Sleep(20000);
				}
			}
			catch (Exception ex)
			{
				Debug.WriteLine(ex.Message);
			}
		}

#if SERIAL_ASYNC_READ
		private static void SerialDevice_DataReceived(object sender, SerialDataReceivedEventArgs e)
		{
			SerialPort serialPort = (SerialPort)sender;

			switch (e.EventType)
			{
				case SerialData.Chars:
					string response = serialPort.ReadExisting();

					Debug.WriteLine($"RX:{response.Trim()} bytes:{response.Length}");
					break;

				case SerialData.Eof:
					Debug.WriteLine("RX :EoF");
					break;
				default:
					Debug.Assert(false, $"e.EventType {e.EventType} unknown");
					break;
			}
		}
#endif
	}
}

When I requested the RAK3172 version information with the AT+VER? command the response was three lines, consisting of the version information, a blank line, then the result of the command. If I sent an invalid command the response was two lines, a blank line then “AT_ERROR”

AT+VER? response synchronous

The asynchronous version of the application displays character(s) as they arrive so a response could be split across multiple SerialDataReceived events

AT+VER? response asynchronous

Unlike the RAK811 module the RAK3172 defaults 9600 baud which means there is no need to change the baudrate before using the device. I use the excellent RaspberryDebugger to download application and debug them on my Raspberry PI 3.

SeeedLoRaE5-NetCore on Github

The source code of my .Net Core C# library for Seeed LoRa-E5 modules used in products like the LoRa-E5 Development Kit, LoRa-E5 mini and Grove-LoRa-E5 is now available on GitHub.

A sample application which shows how to connect using Over the Air Activation(OTAA) or Activation By Personalisation(ABP) then send and receive byte array/Binary Coded Decimal(BCD) messages .

//---------------------------------------------------------------------------------
// Copyright (c) September 2021, devMobile Software
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// Must have one of following options defined in the nfproj file
//    PAYLOAD_BCD or PAYLOAD_BYTES
//    OTAA or ABP
//
// Optional definitions
//    CONFIRMED For confirmed messages
//    RESET for return device to factory settings
//
//---------------------------------------------------------------------------------
namespace devMobile.IoT.NetCore.SeeedE5LoRa.LoRaWanDeviceClient
{
   using System;
   using System.Diagnostics;
   using System.Threading;

   using devMobile.IoT.LoRaWan;

   public class Program
   {
      private const string SerialPortId = "/dev/ttyS0";
      private const string Region = "AS923";
      private static readonly TimeSpan JoinTimeOut = new TimeSpan(0, 0, 20);
      private static readonly TimeSpan SendTimeout = new TimeSpan(0, 0, 15);

      private const byte MessagePort = 15;

#if PAYLOAD_BCD
      private const string PayloadBcd = "010203040506070809";
#endif
#if PAYLOAD_BYTES
      private static readonly byte[] PayloadBytes = { 0x09, 0x08, 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01 };
#endif

      public static void Main()
      {
         Result result;

         Debug.WriteLine("devMobile.IoT.SeeedE5LoRaWANDeviceClient starting");

         try
         {
            using (SeeedE5LoRaWANDevice device = new SeeedE5LoRaWANDevice())
            {
               result = device.Initialise(SerialPortId, 9600);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"Initialise failed {result}");
                  return;
               }

#if CONFIRMED
               device.OnMessageConfirmation += OnMessageConfirmationHandler;
#endif
               device.OnReceiveMessage += OnReceiveMessageHandler;
#if RESET
               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Reset");
               result = device.Reset();
               if (result != Result.Success)
               {
                  Debug.WriteLine($"Reset failed {result}");
                  return;
               }
#endif

               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Region {Region}");
               result = device.Region(Region);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"Region failed {result}");
                  return;
               }

               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} ADR On");
               result = device.AdrOn();
               if (result != Result.Success)
               {
                  Debug.WriteLine($"ADR on failed {result}");
                  return;
               }

               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Port {MessagePort}");
               result = device.Port(MessagePort);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"Port on failed {result}");
                  return;
               }

#if OTAA
               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} OTAA");
               result = device.OtaaInitialise(Config.AppEui, Config.AppKey);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"OTAA Initialise failed {result}");
                  return;
               }
#endif

#if ABP
               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} ABP");
               result = device.AbpInitialise(Config.DevAddress, Config.NwksKey, Config.AppsKey);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"ABP Initialise failed {result}");
                  return;
               }
#endif

               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Join start Timeout:{JoinTimeOut.TotalSeconds} Seconds");
               result = device.Join(true, JoinTimeOut);
               if (result != Result.Success)
               {
                  Debug.WriteLine($"Join failed {result}");
                  return;
               }
               Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Join finish");

               while (true)
               {
#if PAYLOAD_BCD
                  Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Send Timeout:{SendTimeout.TotalSeconds} Seconds payload BCD:{PayloadBcd}");
#if CONFIRMED
                  result = device.Send(PayloadBcd, true, SendTimeout);
#else
                  result = device.Send(PayloadBcd, false, SendTimeout);
#endif
#endif

#if PAYLOAD_BYTES
                  Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Send Timeout:{SendTimeout.TotalSeconds} Seconds payload Bytes:{BitConverter.ToString(PayloadBytes)}");
#if CONFIRMED
                  result = device.Send(PayloadBytes, true, SendTimeout);
#else
                  result = device.Send(PayloadBytes, false, SendTimeout);
#endif
#endif
                  if (result != Result.Success)
                  {
                     Debug.WriteLine($"Send failed {result}");
                  }

#if LOW_POWER
                  Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Sleep");
                  result = device.Sleep();
                  if (result != Result.Success)
                  {
                     Debug.WriteLine($"Sleep failed {result}");
                     return;
                  }
#endif

                  Thread.Sleep(60000);

#if LOW_POWER
                  Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Wakeup");
                  result = device.Wakeup();
                  if (result != Result.Success)
                  {
                     Debug.WriteLine($"Wakeup failed {result}");
                     return;
                  }
#endif
               }
            }
         }
         catch (Exception ex)
         {
            Debug.WriteLine(ex.Message);
         }
      }

#if CONFIRMED
      static void OnMessageConfirmationHandler(int rssi, double snr)
      {
         Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Send Confirm RSSI:{rssi} SNR:{snr}");
      }
#endif

      static void OnReceiveMessageHandler(int port, int rssi, double snr, string payloadBcd)
      {
         byte[] payloadBytes = SeeedE5LoRaWANDevice.BcdToByes(payloadBcd);

         Debug.WriteLine($"{DateTime.UtcNow:hh:mm:ss} Receive Message RSSI:{rssi} SNR:{snr} Port:{port} Payload:{payloadBcd} PayLoadBytes:{BitConverter.ToString(payloadBytes)}");
      }
   }
}