Student Projects
Contact Information
Competition Year: 2016
University: University of Southampton
Team Members (with year of graduation): Radovan Gallo (2017)
Faculty Advisers: Dr. Dina Laila
Email Address: rg2g12@soton.ac.uk
Country: UK
Project Information
Title: Development of a Low-Cost SCARA Robotic Manipulator
Description:
The aim of this project is to design, build and test a prototype of a low-cost and desktop-sized robotic manipulator. The system will be easy to integrate into non-industrial settings and allow quick task changes thanks to replaceable end effectors.
Products: NI myRIO-1900, LabVIEW 2015, LabVIEW myRIO Toolkit
Three NEMA 17 bipolar stepper motors, three DRV8825 stepper motor drivers, 24V power supply and microswitches
The Challenge:
Robotic manipulators are commonly used in assembly line environments to increase productivity, improve quality and promote safety. However, their uptake for small and medium sized business has been lacking due to their high cost, environmental requirements and limited capability to coexist with the human workforce. Moreover, due to the high cost of retooling, the flexibility of tasks these systems perform is limited. Recently, several companies have started developing robotic systems that can be integrated into existing environments more easily and allow greater flexibility. In this project, a desktop based robotic manipulator is developed that can automate repetitive tasks and can be easily used in an office, studio or laboratory environment.
The Solution:
A SCARA configuration was chosen as it offers the dexterity necessary to complete majority of the tasks, while requiring only 3 actuators, which also reduces the cost. The two planar links are not compliant in the vertical direction, therefore vertical forces created by the end effector are carried by the structure rather than actuators.
The mechanical structure of the robotic manipulator was designed to be manufactured using available rapid prototyping tools. The main structure of the robot is laser-cut from a sheet of PMMA, with other miscellaneous parts manufactured on a 3D printer. To achieve high accuracy, the joints and bearing housings were precisely machined on a metal lathe out of aluminium bar stock.
The end effectors are easily replaceable thanks to the magnetic mounting system. A circular opening locates the end effector in-plane and strong neodymium magnets locate and hold it vertically. This allows different end effectors to be swapped and different tasks to be performed.
For the sake of clarity, the entire structure of the robot was hierarchically grouped and split into four main sub-assemblies, so that they could be worked on independently. These subassemblies are the base of the robot, the shoulder joint, the first planar link and the second planar link. The main dimensions of each of these groups are parametrically controlled and can be adjusted with ease. The workspace of the robot covers an area of an A3 sheet of paper.
(a) Isometric view and (b) side view of the overall robot assembly
Using commercial off-the-shelf (COTS) components wherever possible has allowed to keep the overall cost within budget. The motion and power are transmitted using T2.5 timing belts and pulleys that are commonly used for home built 3D printers. The ball bearings used are standard 608 sized bearings that are commonly used for roller skates.
The robotic manipulator is controlled using the National Instruments myRIO-1900 embedded device.It features a powerful dual-core ARM CortexA9 real-time processor, coupled together with a Xilinx Zynq 7010 FPGA for I/O operations. Thanks to its 88 reconfigurable analog and digital I/O it can easily control the three NEMA17 bipolar stepper motors in an open-loop fashion. The stepper motor drivers are modules based on the DRV8825 stepper motor driver integrated circuit and are capable of up to 1/32-step microstepping.
An illustration of the electronics subsystem
The control software is implemented in National Instruments LabVIEW 2015 graphical programming environment. The advantage is that LabVIEW is inherently capable of parallel code execution and the modular organisation of programmes into SubVI promotes easy code reuse. The motor control routine is implemented using a state machine that generates a trapezoidal velocity profile that the motors follow. The design of the state machine is based on the application note AVR446 by Atmel. By utilizing the FPGA on the myRIO device, all three motors can be controlled in parallel and the timing of the control pulses accurately generated. Microswitches are used for each degree of freedom to enable initial calibration and end-of-travel detection.
The entire system was successfully designed, built and tested over a period of 7 months. The robotic manipulator was tested with a mass of 0.2 kg to simulate an end effector with angular accelerations of up to 7.00 rad.s−2.
While the project has been successful in developing a working low-cost robotic manipulator, several limitations were encountered during the project that could be addressed in the future. First of all, the mechanical structure of the robot could be optimised using finite element analysis. The optimised structure would benefit from lower mass, inertia and improved material usage.
Secondly, the use of other manufacturing techniques and materials could be investigated. Utilising high-performance materials, such as carbon fibre composites, would decrease the moment of inertia of each link and lead to better performance of the system. However, many of these materials are prohibitively expensive and care needs to be taken to weigh the cost against the performance of the system.
Moreover, the greatest drawback of the motion system is the lack of feedback transducers that would enable closed-loop operation. Having feedback sensors mounted on the robot, such as rotary encoders or inertial measurement units, would increase the positioning accuracy and enable a thorough testing procedure that would assess the accuracy, repeatability and reliability of the entire system.
The current control scheme allows the robot to navigate through waypoints in a point-to-point fashion. This approach enables many application, including pick-and-place operations. However, additional applications would benefit from a path following control algorithm. This would enable the end effector follow a given path precisely, which would open the possibility for the manipulator to be used for tasks such as 3D printing, adhesive dispensing, or pen marking.
Nominate Your Professor (optional)
I would like express my deep gratitude to my supervisor Dr Dina Shona Laila for the endless support, guidance and motivation she has offered me during this project. She is the most committed and hard-working academic I have ever encountered and without her deep insights and dedication this project would have never been possible. Her vast experience of using LabVIEW and myRIO has allowed me to understand the development process in greater detail and made using National Instruments products that much easier.
