VCS – Vision, Control and Sensor solution for Precision Robotics

Aug 9, 2019

The VCS-1 is a PC/104 Linux stack composed of 2 main components, namely the EMC2 board which is a PCIe/104 OneBank™ carrier for a Trenz compatible SoC Module and the FM191 expansion card that fans out the I/Os from the SoC to the outside world.

The VCS-1 system is a complete and easy-to-use solution comprising of both software and hardware, it was designed and developed in the VineScout project to collect data from state-of-the-art cameras.  It is compatible with a wide range of sensors and actuators.

Sundance VCS-1 controlling the SUMMIT-XL

The video below demonstrates the Sundance VCS-1 controlling the autonomous navigation of the Robotnik SUMMIT-XL whilst simultaneously collecting images from the two Intel RealSense D435i cameras.
 
The integration of the VCS-1 into the SUMMIT-XL robot took around 4 man-months, but Sundance is confident that we could port the VCS-1 system to any robot within a man-month to provide similar autonomous navigation and provide NDVI and other measurements using the Deep Learning and Edge-AI integrated into the VCS platform.

Development of the SUMMIT-XL control project

Creating the mounting points for the VCS-1 in the front of the Summit-XL gives easy access and therefore ensures that maintenance of the system is simple.  The use of a USB 3.0 hub means that implementing external devices such as the GPS is also easy.
 
Wiring the power supply directly to the terminal with the power switch connected to the robots peripherals on off switch enables the VCS-1 to be powered separately from the robot itself.
 
Initially the cameras were mounted to the SUMMIT-XL using the tripods they are supplied with.  This gave an opportunity to fine tune the position of the cameras to best collect the data.  A 3D printed frame was then developed using the experience gained testing with the tripods, allowing a dynamic and adjustable frame to be designed and manufactured.
 
Initial field tests using the VCS-1 to autonomously navigate the SUMMIT-XL using distance were successful.  This then allowed the Summit to be able to collect data and navigate through a set course simultaneously.

Intel® RealSense™  D435i

These D435i cameras combine the robust depth sensing capabilities of the D435 with the addition of an inertial measurement unit (IMU).  An IMU combines a variety of sensors with gyroscopes to detect both rotation and movement in 3 axes, as well as pitch, yaw and roll.

Adding an IMU allows your application to refine its depth awareness in any situation where the camera moves. This opens the door for rudimentary SLAM and tracking applications allowing better point-cloud alignment. It also allows improved environmental awareness for robotics and drones. The use of an IMU makes registration and calibration easier for handheld scanning system use cases and is also important in fields such as virtual/augmented reality and drones. When using the D435i, our Intel RealSense SDK 2.0 provides IMU data that is time stamped to align with our high quality depth data.

Robotnik SUMMIT-XL

The SUMMIT-XL has skid-steering kinematics based on 4 high power motorwheels. Each wheel integrates a hub brushless motor with gearbox and encoder (optional).

SUMMIT-XL has two possible kinematic configurations. The omnidirectional configuration mounts mecanum wheels on an independent suspension system. The mecanum wheels can be easily replaced by conventional wheels (rim mount), thus allowing easy switch from the indoor omnidirectional configuration to the versatile skid-steering configuration, both indoors and outdoors.

The odometry is computed with the use of the four encoders and a high precision angular sensor mounted inside the chassis.  The strong mechanical structure allows to carry high loads. There are several suspension shocks possibilities. They can also be mounted at several positions to modify the robot clearance.  The robot base can navigate autonomously or teleoperated by means of a PTZ camera that transmits video in real time.  The common sensor options include a Hokuyo laser scanner and a range of RTK-DGPS kits. It also has internal (USB; RS232 and GPIO) and external connectivity (USB, RJ45 and power supply 12 VDC) to easily add custom components.

The control architecture is open-source and modular, based in ROS