Robotik-Controller with VPU/GPU cards

Roscube robotic controller series support customers implementing their individual AI application

Artificial intelligence, especially machine learning, is finally also attracting increasing interest in more traditional industrial applications. Be it in maintenance and error forecasting or the optimisation of operating and production parameters: the use of specialised and powerful hardware effectively supports the rapid development of software and algorithms. Acceed has now added several robotic controllers of the ROScube series to its product range in order to optimally support these development steps of its customers. The ROScube-X has an integrated Jetson module, while the ROScube-I is specifically designed for expansion with Intel VPU or Nvidia GPU.

Jetson is the brand name of the manufacturer Nvidia for a series of integrated computer chips with Tegra processor. Jetson is a low power system specifically designed to accelerate machine learning (ML) applications. Machine Learning, also called machine learning in the German language, can be considered a sub-area of Artificial Intelligence (AI). It uses algorithms to build a mathematical model based on so-called training data in order to make predictions or decisions without being explicitly programmed for it. Machine learning algorithms are increasingly used in industrial applications, for example for the diagnosis of machines and plants, for the optimisation of settings during the commissioning phase or for the autonomous adaptation of processes during operation. Even the development of new products can be significantly accelerated by using machine learning. Ultimately, the aim is to increase efficiency, productivity and availability, which cannot be achieved with conventional programming and simple control.

With the Jetson Xavier NX, Nvidia introduced a new developer board for the machine learning sector last year. The board was developed for use in robots and edge computing, among other things, and is used in the robotic controllers ROScube-I and ROScube-X, which the German distributor Acceed has now added to its range.

The ROS2-capable ROScube-I is optionally equipped with different Intel processors i7, i3 or i5-BGA of the 9th generation. The interfaces, which are especially designed for use in robotics, serve to connect a large number of sensors and actuators, thus providing the greatest possible scope for implementing high-performance robotics applications. An optional expansion box (PCIe Gen 3) also allows the installation of application-specific hardware extensions. The ROScube-I is optimised for use with Intel VPU cards or Nvidia GPU cards to calculate AI algorithms and inferences. The robotics controller supports all developments with the Neuron SDK from Adlink for industrial use, for example autonomous mobile robots (AMR) and autonomous mobile industrial robots (AMIR).

The ROScube-X contains the Jetson-Xavier module and also an integrated Volta-GPU, both from Nvidia, as well as two deep learning acceleration units. For development purposes, the complete range of functions of the JetPack-SDK and the Neuron-SDK is supported. This makes the controller an excellent platform especially for robotics applications with high demands on AI computing performance and low power consumption.

Other features include robust, lockable USB ports, a wide range of I/O interfaces, time synchronization with GMSL2 cameras and IMU, remote power-on for robots and real-time access for CAN bus, GPIO and serial interface. The maximum permissible ambient temperature for operation is specified as -20 to +70 °C, depending on the equipment. Further technical details and information on application possibilities are available directly from the German distributor Acceed.