Luna – MANSEDS Lunar Rover
The mission of this team is to build a rover capable of winning the UKSEDS Lunar Rover Competition. Throughout this process we hope to individually discover and learn more about our respective interests, as well as collectively build an ambitious yet achievable rover.
In light of this, we hope to design a lunar rover capable of meeting all requirements and exceeding all expectations, while creating opportunities for ourselves to expand our knowledge in both the theory as well as the practical work associated with complex machines. Ultimately, building a lunar rover presents us with the perfect opportunity to grasp this mission.
This rover should be capable of remotely navigating craters and uneven terrain, collecting ice and rock samples and returning these samples to a lunar lander module. To add a twist, it must be able to do this after enduring a vibration test.
The arm has been designed as a 4 link robotic arm offering 6 DOF. The links have been designed for additive manufacturing, excluding the base rotation kit which has been bought pre-built. The end effector is being designed in the style of an excavator bucket. This end effector will also provide the mounting for a FLIR sensor.
The arm will be actuated by a series of servo motors mounted within the arm joints. Towards the base of the arm, larger torques will be experienced. As such, the servos powering the end effector are the weakest and the servos increase in capability approaching the base. These servos are controlled by the control systems.
Mounted over the front axle of the rover, the arm is capable of reaching at least 150mm beyond any surface of the rover with more reach at the front.
System on a Chip
The control systems utilise the Raspberry Pi system on a chip (SoC). This SoC has a powerful quad-core 1.2GHz ARMv8 processing unit, a graphics processing unit, general purpose input and output, onboard wireless communication hardware and allows for development in C++ & python.
The SoC will function as the brain of the rover. It will be monitoring and controlling all sensors, itself and the electronic actuators.
We will use the RPi’s onboard WiFi chipset to provide the communications necessary to control the rover remotely. Utilising the Linux OS, a web server will be created on the RPi that will allow for both the human controller to send instructions to the rover, as well as for the rover to send appropriate information – such as video feeds and orientation data – to the human controller.
Code to monitor sensors, stream data to an operator and control all actuators will be running on the SoC. To control the arm, an inverse kinematics scheme is being developed. The drive system will be easily directed through a simple set of instructions that allow an operator to remotely navigate through rough terrain. This code also allows for the panning and positioning of cameras.
Ultimately, a good human/machine interface should be intuitive and a child should find it easy to operate the vehicle remotely. This interface will consist of a web application providing camera feeds and telemetry streams allowing the operator to act on the current state of the rover. The control panel is being designed to be easily manageable while not limiting the level of control available.
Two types of camera provide three image streams to the operator interface. Two 2MP USB cameras that are capable of recording videos, with night vision and infrared capabilities. These cameras have sufficient resolution for the pilot to clearly view the rover surroundings in addition to being easily interfaced through two of the SoC’s USB ports. In addition, a FLIR sensor mounted on the arm provides a thermal imaging stream for more easily directing the end effector towards ice samples.
A gyroscope and accelerometer chip will be connected over i2c and will be controlled using the SMBus module for Python. I2c supports up to 100kbps speeds in standard mode, and up to 400kbps speeds in fast mode, which for the 128bits of information that can be acquired from the chip yields polling rates of ~780Hz and ~3100Hz respectively. The gyroscope and accelerometer chip supports both modes so, while in practicality it is likely both data rates are sufficient, we have extra flexibility dependent on empirical data once the system is running. The data pulled from this chip will be both sent to the human controller as well as analysed onboard to make automatic emergency stop decisions if destabilisation is detected.
The proximity sensors will be connected over GPIO pins, and so we will use the provided GPIO module of the RPi library for Python. We are in the process of testing a variety of proximity sensors to determine which sensors are most appropriate for this vehicle. Unfortunately ultrasound would be ineffective in a lunar rover and as such are not suitable.The data pins for the shortest range sensors will be set up to act as interrupt signals, ensuring accidental collisions are avoided by effecting an emergency stop when triggered. The operator will be able to override this. All data pins will be polled to also provide data to the human controller.
The motors that drive this rover onward are high torque Actobotics motors. The decision was made to increase torque at the cost of speed to allow this rover to power up steep gradients such as may be found in craters.
The motors are controlled electronically with the help of dual H motor bridges. These bridges control velocity in response to a PWM signal. They also determine direction of rotation in response to 2 hilo signals from the SoC GPIO pins. These pins can also send a brake signal which immediately stops the motors.
Mounted directly on the motor shafts, large diameter off-road tyres provide a high ground clearance and maintain grip on unfavourable surfaces.
Aluminium Profile Struts
A three tiered structure housing all components except the arm and motors will be constructed from aluminium profile struts. These struts provide rigidity while reducing weight. They also provide a wide range of mounting options which makes mounting all internal components simple and sturdy.
To ensure that the vibration tests do not shake the structure apart, nylock nuts will be used wherever feasible. T-bolts will be used to fasten components to the struts.
To provide protection from the elements and space for sponsor logos, fairings will surround the structure. These fairings will be cut from correx; this provides a waterproof, lightweight set of fairings which can be easily branded.
Lunar Rover Project 2017
The Lunar Rover project is an endeavour to design and build a rover capable of being sent to and operated on the Moon. The project was started in 2016 and saw an interdisciplinary group of engineers and physicists come together and tackle the problem of collecting ice samples on the lunar surface. In the coming year, we hope to engage a new wave of students with the project and improve the design in preparation for the 2018 Lunar Rover Competition by UKSEDS. We will be rebuilding the rover from the ground up, so this is an excellent opportunity to work on a robotics project and gain lots of experience.
The project will be split into several stages: initially we will investigate the flaws with the previous year’s design and what changes we can make to fix those flaws. We will then begin producing a preliminary design report, which will be critiqued in time for the competition and once we’re happy with the design we’ll begin construction. All the while the intent is to provide classes on related topics, such as programming, CAD work and so on.
If you are interested by the idea of joining an awesome project where you get to play with electronics, robotics, structural design, code and all manner of other cool things, then please get in contact! You can do that by messaging any of Matthew Marshall, Derya Kirdag or Ethan Ramsay. Alternatively, if you prefer email, you can contact the lunar rover email.