BlackBerry QNX Academy for Functional Safety

###<header>Putting Together the QNX AI Robot Demos</header> It's time to create some fun demos with the robot. <hr/> ⚠ Make sure to obtain the QNX evaluation licenses first as the QNX Foundry and the QNX GitLab are closed sites and are only available once you have the licenses. <hr/> ####<header2>Installing the QNX Pre-Built AI Image onto a microSD:</header2> Make sure the QNX OS for Safety 2.2.1, QNX SDP 7.1, and Hypervisor for Safety 2.2 are installed. Now you will be able to access and download the QNX pre-built AI image, which will contain all of the robot demo code and patches needed in this Academy. Log into your GitLab account (with the same email used for your myQNX account) and follow the <a href="https://gitlab.com/qnx/external-collaborations/blackberry-qnx-ti-academy/blackberry-qnx-ti-academy-docs" target="_blank">QNX pre-built AI installation instructions</a> to install the image onto a microSD card. ###<header>Running the QNX AI Robot Demos</header> To run the demos, follow the <a href="https://gitlab.com/qnx/external-collaborations/blackberry-qnx-ti-academy/blackberry-qnx-ti-academy-docs" target="_blank">QNX AI robot demos instructions</a>. You will build and run the following robot demos in stages: <ul> <li>startup verifier and fault injector</li> <li>object detection and ranging</li> <li>subject follower</li> <li>LiDAR-based navigation</li> </ul> ####<header2>Startup Verifier and Fault Injector</header2> The startup verifier checks the system startup configurations are correct before running the system. If the startup configurations passes without any failures, the startup verifier runs the QNX OS for Safety. The fault injector tests the operations of the startup verifier, whereby a fault is injected. If a fault is found by the startup verifier, the system will move to a design safe state where the system is prevented from taking any form of coherent action. ####<header2>Object Detection and Ranging</header2> For the first demo, the robot needs object categorization and depth perception. Vision convolution neural network (Vision CNN) runs a deep learning algorithm that simply makes use of a camera to detect different objects in the scene and categorizes them. <img src="./resources/obj_det_example2.gif" width="840" height="480" class="card-img-top" alt="..."/> Stereo depth estimation (SDE) is a technique of using two cameras and using epipolar geometry to calculate the depth of different objects in a scene. This technique requires significant computing resources. The highly integrated TDA4VM edge AI processor has a special hardware accelerator for SDE which is optimized for high-performance, real-time computer vision algorithms such as localization, mapping, scene analysis, and object detection in automotive and robotics. <img src="./resources/5_demo.gif" width="840" height="480" class="card-img-top" alt="..."/> Combining vision CNN and SDE will give the robot object detection and ranging capabilities. ####<header2>Subject Follower</header2> Next, add motor code to the robot with the object detection. Using the object detection, make the robot choose a specific subject and follow it while still detecting the other objects. In this case, it could be a basketball. <img src="./resources/4_demo1.gif" width="840" height="480" class="card-img-top" alt="..."/> ####<header2>LiDAR-based Navigation</header2> The next step is to add navigation capability to the robot. The robot needs to know the map of the environment, have the capability to know its own location, and determine the path to the destination. <img src="./resources/0_demo3.gif" width="840" height="480" class="card-img-top" alt="..."/> Putting it all together, you will have built an autonomous robot. <hr/> ###<header2>Next Steps</header2> After going through the BlackBerry QNX Academy for Functional Safety and completing your robot demos, let's summarize what you've learned. ##### [Summary of the BlackBerry QNX Academy for Functional Safety >](4_summary.html) <hr/> <div align="center" style="margin-top: 4em; font-size: smaller;"> © 2022 BlackBerry Limited. All rights reserved. <a rel="license" href="https://creativecommons.org/licenses/by-nc-nd/4.0/"> <img alt="Creative Commons License" style="border-width:0" src="..//web_support/cc_license_icon.png" /> </a> This work is licensed under a <a rel="license" href="https://creativecommons.org/licenses/by-nc-nd/4.0/"> Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License </a>. </div>