Getting Started With Jetson Nano

This section describes how to run Isaac SDK sample applications on the Jetson Nano device. For directions on how to get started with Nano in general, see Getting Started with the Jetson Nano Developer Kit.

Obtain the IP address of Jetson Nano:

1. Connect a keyboard, mouse, and display, and boot the device as shown in the Setup and First Boot section of Getting Started with the Jetson Nano Developer Kit.

2. At a terminal prompt, enter the following command:

   Copy

   Copied!

               
   
               
   bob@jetson:~/$ ip addr show
           

This section describes the steps to run sample applications on Jetson Nano. The first sample does not require any peripherals. The second sample is a more useful application that requires a connected camera. The third sample demonstrates how to deploy a TensorFlow model and run inference on the device.

Other applications can be deployed and run using the methods described here.

Ping

1. Deploy //apps/tutorials/ping:ping-pkg to the robot as explained in Application Console Options.

2. Change to the directory on your Jetson Nano and run the application with the following commands:

            

            
bob@jetson:~/$ cd deploy/<bob>/ping-pkg/
bob@jetson:~/deploy/<bob>/ping-pkg$ ./apps/tutorials/ping/ping
        

Where <bob> is your username on the host system. You should see “Hello World!” print every 1.5 seconds.

OpenCV Edge Detection

1. For this sample, connect a camera to one of the USB ports on the Jetson Nano.

2. Deploy //apps/tutorials/opencv_edge_detection:opencv_edge_detection-pkg to the robot as explained in Application Console Options.

3. Change to the directory on your Jetson Nano and run the application with the following commands:

            

            
bob@jetson:~/$ cd deploy/<bob>/opencv_edge_detection-pkg/
bob@jetson:~/deploy/<bob>/opencv_edge_detection-pkg/$ ./apps/tutorials/opencv_edge_detection/opencv_edge_detection
        

4. To view the results, load http://<nano_ip>:3000/ in your browser. Make sure that the application is running when you are loading the webpage.

TensorFlow Model Inference with TensorRT

1. Deploy //apps/samples/ball_segmentation:inference_tensorrt-pkg to the robot as explained in Application Console Options.

2. Change to the directory on your Jetson Nano and run the application with the following commands:

            

            
bob@jetson:~/$ cd deploy/<bob>/inference_tensorrt-pkg/
bob@jetson:~/deploy/<bob>/inference_tensorrt-pkg/$ ./apps/samples/ball_segmentation/inference_tensorrt
        

3. To view the results, load http://<nano_ip>:3000/ in your browser. Make sure that the application is running when you are loading the webpage.

How to Flash Jetson Nano

Nano can be flashed in two ways:

• Via an SD Card: See the procedures in Getting Started with the Jetson Nano Developer Kit.

• Via SDK Manager: See the procedures in SDK Manager User Guide

Use the documents above to determine the best method for your use case.

How to Connect a Wi-Fi/Bluetooth Card

The Intel 8265 card is used for Wi-Fi and Bluetooth connectivity. The following steps describe how to install a Wi-Fi/Bluetooth card for Jetson Nano.

1. To access the M.2 slot on the carrier board, remove the two screws on the side and open the SODIMM latches using both your hands.

   

2. When the Jetson Nano module pops up, slide it out gently.

   

3. Take out the Intel wireless card and attach the antenna to its U.FL sockets before inserting the card into the M.2 socket. Attaching the antenna on the sockets requires patience and getting used to. Use your nail to gently apply force. You do not need much force to clamp it in once you are in the right position.

   

4. Slide the Intel 8265 card into the socket.

   

5. Fix the Intel 8265 card in place with a screw, and replace the Jetson Nano module. Make sure to use the correct screws in each case.