Autonomous vehicle software testing and integration

Eric took a role in the Corporate R&D group focusing on the development, integration and test of Autonomous vehicles (ground and air). The robots used advanced mapping techniques to self-navigate the environment. Eric worked with highly accurate sensors; Time of Flight cameras, Stereo cameras, Sonar, IMU data, April Tags, etc. for autonomous navigation using different methodologies; SLAM, VISLAM, Optic Flow, VIO, RoS VIO. Eric used the RoS (robotic operating system) to test autonomous software packages on the robot and in simulation (Gazebo). Eric worked directly with the Autonomous source code, editing launch files for RoS, reading C++ and python code. Eric compiled source code of ROS Indigo packages using catkin_make for different autonomous robots, i.e. kobuki, turtlebot, etc. Eric created test content exercising the machine vision software and autonomy packages that were under development. These critical components were automatically tested when a change to the software build was merged by the development team. Eric worked with development tools, such as Git, Github and Gitlab to view and edit source code. Eric contributed to the Autonomy build Git repo by merging files for specific robots, i.e. yaml files, configuration files and updating documentation. These contributions removed several manual steps in the setup process and enabled quick testing activities focused on autonomy.

As a lead validation engineer, Eric tested and successfully launched new products by playing a key role in the integration of cell phone SoC’s into different IoT applications, specifically autonomous vehicles or drones (Snapdragon 821, 801, etc.). Eric worked closely with development team on complex features for mapping the environment; such as using camera technology for mechanical flight control; VIO, VSLAM, VISLAM, Optic Flow. Worked with in-house developed Machine Vision (MV) software library enabling visual obstacle avoidance during flight (VoA). Performed camera calibration for stereo cameras giving 6 DoF and monocular tracking camera to improve localization accuracy. Uniquely, the process which controlled the flight stack ran on the DSP processor and not the applications processor. Eric assisted in profiling the performance of the DSP using sysMon, a tool from the Qualcomm Hexagon tool kit. Eric also assisted in power profiling high power uses cases on the drone platform. Eric provided documentation to enable 3 camera concurrency during flight, i.e. FPV, Video encode with digital gimbal and depth from Stereo (VoA). Eric further performed deep debugging on logs collected during test using tools like; QXDM, QPST, JTAG, QFIL, matlab, git, python, ROS, RosViz, Linux sh, adb, dmesg, logcat & motion capture software. The use of these deep diving tools enabled quick root cause and action plans for a fix. Eric also supported customer demonstrations by flying the vehicle manually and presenting its capabilities. Eric displayed key features of SoC running proprietary software called SNAV (Snap Dragon Navigator), these capabilities included; visual obstacle avoidance, hover accuracy, GPS flight capability, and VIO flight mode stability were all shown during the demo. Lastly, Eric created over 500 test and use cases to fully validate the Visual Obstacle Avoidance feature.

As a Power test lead was responsible for facilitating and delivering system level power consumption values on key uses cases for Qualcomm’s SoC’s, specifically Snapdragon 810, 808, 805 & 801. Focus was to have the best power/performance and user experience. Debugged problem areas by analyzing use case power consumption waveforms, CPU cluster residencies, CPU and GPU Clock rates, DDR Bandwidth consumption, MIPS, FPS, etc. Extensive use of java automation for quick data collection enabled round the clock power data collection possible. Developed new automation as needed for customer uses cases i.e., an automated use case to profile specific games (angry birds, temple run, etc), then tune the system to give the best user experience.

Delivered detailed platform power measurement on the Atom based SoC x86 Tablet. Deep debug analysis on platform power with a focus on optimizing SoC power, CPU and GPU. Other platform optimization activities included analyzing power consumption for backlight, panel electronics, touch, WiFi, Bluetooth and other important platform power rails.

Represented Validation engineering’s input into the Product Development Team by providing key validation data identifying risks and impacts on go/no-go decision for product launch. Managed Validation testing resources delivering on time requirements to enable a successful launch of multiple Wireless (WiFi) Sustaining and New Product Initiative (WiFi / BT) software products supporting 22 different hardware Sku’s. Validation areas include:• WHQL Certification Team• Customer Performance Collaterals• FCC Regulatory Testing• WiFi Alliance Certification• Bug database management• Software Quality Release Criteria

Selected to drive validation engagement for innovative technologies developed to provide wireless synching / enhancement for multiple media technologies. Serve as acting team lead with ownership of information and processes to strategize critical time to market project stages. Oversee and generate WiFi High Level Plans; develop / execute test plans and reviews, capacity planning, and reporting. Manage bug verification, updates, posts, releases, and announcements from testing; report Quality Center indicators. Modify and adapt software; create and automate test cases to upgrade interfaces and enhance performance. Communicate with customers and teams to develop specifications, requirements, and align interface with client needs. Report status, product stages, and health to stakeholders and management.