Data transmission from LiDAR, syslog

[mkonwerski]

Hi. I don’t understand how the LiDAR and SYSLOG data are transmitted to the computer via ESP32 Wi-Fi. In the config file, the desktop computer’s IP and port are set, which suggests that communication is done through micro-ROS. I started the micro-ROS bridge on port 8889, but the ESP32 did not connect to that port.

Could someone explain to me how LiDAR and syslog communication is carried out?

[hippo5329]

The syslog and Lidar forwarder run on udp. They are independent of micro-ros.

You should start from basic micro-ros reconnection example described in the tools installation wiki. And you should start with serial transport first.

[hippo5329]

https://github.com/hippo5329/micro_ros_arduino_examples_platformio/wiki#micro_ros_arduino_examples_platformio

[mkonwerski]

I have two robots. The one based on the ESP32 with Wi-Fi helped me understand a lot and discover its limitations, especially the issues related to Wi-Fi communication.

[hippo5329]

Have you succeeded with micro-ros, syslog and Lidar?

Most of my robots are based on esp32 wifi. It is essential to have some good and fast 2.4g wifi router/ap. I use tp-link ax21. They work well.

[mkonwerski]

I am at the beginning of the journey; I still need to create a URDF file and perform better tests, but what I have looks promising.
I have a good router, and I managed to drive the robot around the house using lidar data.
The robot is controlled by the General Driver Board for Robots (IMU, MAG, INA, encoder) + LS19. I disabled OAT to avoid clogging the Wi-Fi and to reduce the processor load. I might also disable syslog.

[hippo5329]

You do not need to create a fresh new URDF. You need only update some key parameters.

You may disable the magnetometer and compare navigation result. I2C bus bandwidth is quite limited even at 400KHz clock.

You should use OTA to update firmware after the first serial upload. You do not need to connect a USB cable to do this. It will be easier than move the robot to your computer. It won't add much load to the MPU.

The syslog is very useful for debug messages. It replaces the serial print. The wifi signal strength is monitored. You may insert more syslog in your firmware during development.

[mkonwerski]

My robot is so small that connecting a USB cable is not a problem, so I don't need OTA.

I'm having an issue with micro-ROS restarting, possibly due to a ping problem — is there any way to fix this?
case WAITING_AGENT: EXECUTE_EVERY_N_MS(500, state = (RMW_RET_OK == rmw_uros_ping_agent(100, 1)) ? AGENT_AVAILABLE : WAITING_AGENT;); break;

What is the orientation of the IMU and magnetometer on the Gendrv board?

[hippo5329]

The side with antenna is front and the side with USB connectors is back.

You may try set the orientation of antenna to vertical. You may check syslog. There is wifi RX signal strength.

[PaulBouchier]

@mkonwerski I too have seen the problem of excessive restarting. I hacked around that particular problem by requiring 3 ping drops before restarting:
EXECUTE_EVERY_N_MS(200, state = (RMW_RET_OK == rmw_uros_ping_agent(100, 3)) ? AGENT_CONNECTED : AGENT_DISCONNECTED;);
However, something is wrong in wifi - I see periods where wifi just stops for up to two seconds. This made IMU not work, because the EKF filter would use a stale version of IMU data for that period, which ranged anywhere between 20ms and 2 seconds. This caused loss of localization. Its high on the priority list to look at - I don't think it's a wifi signal issue. If you have any insights, please share them.

[hippo5329]

I did have the restart issue in the early days. It is mostly gone when I switched to tp-link ax21.