How to achive autonomous water surface boat using ZedCam, Rplidar S2 on Jetson Nano TX2?

I want to create an autonomous boat that should be operated on Jetson Nano TX2 (Ubuntu 18.04) , and ZED cam and RPLidar S2 which both should be data fused and used to avoid obstacles on the surface of the water. I am literally confused about where to start. I don’t know how to get the output of the Zed cam’s and RPlidar S2’s in Jetson TX2. When I search for online resources I can’t get a clear output of what I want. Please can anyone help me to achieve the task? I know this is big still, but I have the most interest in creating it. I have all components but still, I can’t do anything with it.

Hi @shrijayan, I would start by installing the ZED SDK for Jetson:

This comes with tools that allow you to view/test the camera, and also programming APIs that allow you to access the camera data and stereo depth data from your own application.

It appears that rpLIDAR also has an SDK available here:

Respected sir, @dusty_nv

I will try this but should I install all these in ROS melodic or directly in TX2 ?

Hi @shrijayan, that depends entirely on if you want your application to be based on ROS or not. Since you are making an autonomous robot, in the longer-term it would seem like a good idea. However there is additional complexity of using ROS initially, and much to learn about it. So first I would start by just testing the sensors standalone via their respective SDKs, and then if want installing ROS and the ROS nodes for the ZED and rpLIDAR.

Respected Sir, @dusty_nv
Thank you so much for your help, and my query is not a one-day process I have just started this step, and in this journey, I will try to complete step by step. Once any thank you for your will again contact you for further queries.

Yours Respectfully

Respected Sir, @dusty_nv

Sir, do you have any idea how to integrate any USB camera with jetbot ?

Hi @shrijayan, I would open a question about it on the JetBot github issues tab. I believe you would change this gstreamer pipeline, but I’m not personally sure: