1 System Description

Autonomous robotics has seen significant growth recently in various industrial applications, and is anticipated to further grow. The International Federation of Robotics (IFR) reported that global sales of industrial robots reached a new record of 387,000 units in 2017. That is an increase of 31 percent compared to the previous year ^[1]. It also estimates that more than three million industrial robots will be in use in factories around the world by 2020 ^[2].

This reference design showcases autonomous robotic system development leveraging hardware and software components from Texas Instruments (TI); specifically, the high-speed, real-time processing with the Sitara™ AM57x processors and accurate sensing with the mmWave radar sensors. This reference design demonstrates the functionality of an embedded robotic system, where point-cloud data from the mmWave sensing is communicated and processed by the Sitara™ AM57x processor. The Sitara™ AM57x processor runs the Robot Operating System (ROS) and is the main processor for overall system control.

With a collection of software libraries and packages, the ROS provides a flexible framework for creating robotic applications. Running ROS on top of Linux is a good starting point of developing robotics software. TI’s Processor SDK Linux provides a fundamental software platform for development, deployment and execution of embedded applications on a TI processor running Linux. To enable creation of robotics software, Processor SDK Linux has included ROS (indigo LTS) through the meta-ros open embedded layer. Leveraging widely adopted open source (ROS and Linux) enables easy tuning, enhancement, and addition of robotics modules to meet requirements of various robotics applications.

It is frequently the case that autonomous robotics applications require sensors to enable interaction with the environment. Various types of sensors can be used, such as cameras, ultrasound, time-of-flight RGB-D cameras, and Lidar sensors. An important type of sensor is the mmWave single chip sensor, which enables unprecedented 3D detection and localization of objects across a wide field of view with high spatial and velocity resolution. Compared to optical or light based techniques, an important advantage of mmWave sensors is immunity to challenging sensing conditions including bright or no lighting, translucent or dark surfaces, and dusty or smokey environment. For software, TI’s mmWave SDK provides foundation components to facilitate end users to develop mmWave radar applications using TI's mmWave sensors.

This reference design builds a fully operational robot navigation system with embedding processing and modular ROS. The single ROS host runs on Sitara and acts as a broker for all the inter-node transactions. The point-cloud radar data from the mmWave node is seamlessly communicated to the AM57x processor for real-time processing. Meanwhile, the AM57x processor controls a Kobuki node, for example, the mobile platform, to navigate and avoid obstacles detected by the mmWave sensor. The AM57x processor also communicates with a Ubuntu Linux box for real-time visualization of the navigation.