Competition Year: 2017
University: University of Southampton
Team Members: George Alexopoulos (Team Leader), Radovan Gallo (2017), Aniket Chopda (2017), Yun Yuan Tan (2017), Neophytos Demetriou (2017), Andreas Hadiraftis (2017), Rahul Nair (2017)
Faculty Advisers: Dr Mohamed Torbati
Email Address: M.M.Torbati@soton.ac.uk
The Pegasus Mars Rover
Pegasus Mars Rover Project Overview
Humans have been searching for life in space for many years and now in our generation, scientists and engineers are looking into how human life could be sustained in foreign planets such as Mars. To find life but to also sustain life in Mars autonomous robotics systems will play a key role. Rovers need to be able to carry numerous tasks such as extreme retrieval and delivery of objects, perform scientific tasks, equipment servicing tasks and to assist astronauts in emergency situations.
Figure 1: Uploading code to the Rover prototype
Team Pegasus, from the University of Southampton, is the first team from the UK to participate in the University Rover Challenge. We are a group of eight final year Mechanical Engineering students equipped with passion, willingness, experience, and are united by the same endeavor which is to travel to the US and compete with teams from all around the world. Our team decided to design and build the next generation Mars Rover that will one day work alongside human explorers in the field.
The overall mechanism of the Pegasus rover is based on our three key attributes – robustness, reliability and simplicity. Our team has adopted an iterative design process that has enabled us to develop and test new design concepts rapidly. We have managed to assemble a preliminary testing prototype of the rover and its robotic arm in December, which has enabled us to evaluate the design thoroughly and apply lessons learned to produce an improved system.
Figure 2 - A CAD view of the initial testing prototype (left) and the final rover design (right)
The robotic arm is a 5 DoF articulated system consisting of a rotating base, two rotary joints powered by linear actuators, and a 2 DoF wrist mechanism. The robotic arm is also designed to be easily removable depending on the nature of the task. The workspace of the arm was derived according to the needs of the Equipment Servicing and Extreme Retrieval Tasks where several design iterations were required to identify the most optimal geometry of the joints and actuators. The design also incorporates an interchangeable end effector which connects to the wrist mechanism allowing it to be changed depending on the nature of the task.
The wrist mechanism, actuated by 2 DC motors and a set of bevel gears, enable the gripper to have additional dexterity making it even easier to adapt to difficult situations such as in the case of the Equipment Servicing tasks.
Figure 3 – (Left) Prototype of Robotic Arm lifting a 7.5kg dumbbell. (Right) – Final Robotic Arm Design
A prototype has been built and tested to validate the range of motion and lifting capabilities of the robotic arm, and modifications have been made to reduce its weight. The robotic arm is in the final stages of manufacturing and simulations of different competition scenarios will be run to train the robotic arm operator.
A four-bar linkage configuration was chosen and was implemented with two distinct actuation mechanisms. The final gripper design consists of a DC motor that drives a worm gear which in turn rotates a pair of worm wheels each of which are attached to the fingers. This arrangement prevents backlash and the fingers are interchangeable to fit the requirements of the task.
Figure 4 - Final Gripper Design
The seamless interaction of software and Rover mechanics is the key to assuring the rover’s success. The software team has continually strived to maintain this harmony between the sub-systems. At all stages, the code was kept as simple as possible to avoid complication and thus ensuring good coding practices. We have decided to use National Instrument’s software; LabVIEW. This was decided to accelerate software development and deployment, by avoiding the complexity of text-based programming. This required all the team members to be trained in LabVIEW FPGA and Real - Time by the team leader who is the most experienced LabVIEW user.
Figure 5 - sbRIO-9627
The NI Single-Board RIO controller that we are using, is excellent for such an application due to its small size, power requirements but also hardware architecture. More specifically, our software will include different layers; the driver, platform, algorithms and user interface which can easily be distributed across the real-time processor, the FPGA and a host PC. The FPGA can handle high-speed components such as motor drivers or sensors, which can run deterministically without taking up processing power. In the real-time processor, we can deterministically address the control code from the platform and algorithm layers and then connect the controller to a Wi-Fi network via a router to control the robot remotely using TCP/IP. To set up the Wi-Fi network we used Ubiquiti hardware.
Figure 6 - Aniket writing LabVIEW code
While coding we have taken a modular approach to meet our design philosophies of simplicity and reliability. In this approach, we created a code for each sub- component and then rigorously tested them. To accelerate our productivity and reach deployment faster, we decided to you the powerful myRIO for the testing of our code modules. We were able to create test code very fast and validate the proper performance of each electronic component.
National Instruments Hardware and Software:
sbRIO - 9626
LabVIEW Robotics Module 2016
Real - Time Module
Other Hardware and Software:
8 x Planetary DC Geared Motors
5 x Dual Channel 10A DC Motor Driver
2 x Linear Actuators
2 x Ubiquity Rocket M2 2.4 GHz Radios
1 x Ubiquity AirMax 2G10 2.4GHz 10dBi Omni Antenna
1x Ubiquity Airmax 2G15 - 120 15dBi Sector Antenna
1 x Tycon Power 12v-14v DC-DC Converter and PoE Injector
2 x EtherNet Cameras
1 x Drill
1 x xBox Controller
Various Sensors (Temperature, Humidity, GPS, IMU)
Nominate Your Professor
We would like to nominate Dr Mohamed Torbati who is acting as the supervisor of this project. We would like to thank him for all the support and guidance that he has provided us with.