Sam Cubero - PhD (USQ), BE Hons Mech (UQ)

RESEARCH
  
  
Summary of past, present & future research projects in Mechatronic Engineering
  
  
PhD Dissertation:  “Force, Compliance and Position Control for a Pneumatic Quadruped Robot”. (approx. 1000 pages)
This project was undertaken during 1994 to 1997 under the supervision of Professor John Billingsley.  I successfully simulated,  designed, manufactured, tested and analyzed the performance of a new kind of novel proportional air valve, called the FPGV (Floating Plate Gas Valve), and created a mathematical model describing its behaviour and performance.  This “world first” design was used to control air flow and regulate pressures on both sides of a pneumatic piston to achieve accurate position, speed or force control, with all joints controlled by “high level” PC software.  These actuators were then used on a 12-degree-of-freedom 4-legged walking robot that I also designed, analysed, built, simulated with 3D graphics software and programmed.  This walking robot, known as the “STIC Insect”, uses 9 computers to successfully control its feet to walk, unaided, over level ground and it employs a novel construction method based on tetrahedral pyramid space frames.  The STIC (Space Truss Integrated Construction) Insect is now on display at the Sydney Powerhouse science museum, Australia.

Examples of past research projects

STIC INSECT (1994-1997):  Completed research on a 4-legged walking robot, new FPGV flow & pressure controllers and a new “Blind Search” inverse kinematics method.  Several new technologies were invented, designed and successfully developed during the course of my PhD research, including the world’s first FPGV (Floating Plate Gas Valve), a dual-valve pressure regulator, a position/speed/stiffness controlled pneumatic actuator, a new “Blind Search” Inverse Kinematics algorithm for controlling any type of serial link robotic arm or manipulator, fully operational 3D simulation and control software, 8 slave microcontrollers linked by a serial communications bus which could control the position and stiffness of each of the robot’s 12 degrees of freedom as commanded by a master PC computer, and a fully operational 4-legged walking robot that could walk over flat ground which was built using a novel “Space Truss Integrated Construction” (STIC) method, where each link is a very strong yet lightweight tetrahedral pyramid.  I am hoping to set up and supervise a PhD project of a new STIC Insect walking robot powered by EMLA linear actuators in the near future.

EMLA (2006):  Electro-Magnetic Linear Actuator: A “world first” all-electric bionic muscle that displays variable stiffness/compliance (or springiness) and high speed precise positioning capability, but shows great potential for being much cheaper to build and more energy efficient to run than lead-screw linear actuators, hydraulic cylinders and pneumatic positioning cylinders for a given mechanical power output.  This EMLA invention is currently operational and the first prototype was successfully built and tested.  Although it worked successfully, the force output of the first prototype was quite low, comparable to a weak pneumatic (air) cylinder actuator able to push 3 kg.f or approximately 30N loads.

ESRA (2006):  (Electric Scooter Robot Arm)  A robotic arm and gripper attachment which can pick up and retrieve products (up to 10kg in weight) at heights up to 3m up and 2m (radially) around a mobile electric scooter.  This device was invented by Dr Sam Cubero and built by final year student Mr Nyan Naung within a period of 10 months and has appeared in several newspaper articles.  ESRA allows elderly, infirm or disabled people to collect products off high shelves in grocery stores, or retrieve heavy items or books off tall bookshelves, where they cannot normally reach.  The robot arm can also be designed for fruit picking jobs.

HYDROBUG (2005):  This is a large 6-legged & 4-wheeled Hydraulic "Bug" robot for transporting 2 human passengers over rough terrain.  The Hydrobug, which I invented in 1999, is designed to walk over large potholes or chasms and place its legs on top of tall obstacles to perform very rough terrain locomotion tasks, eg. Like climbing over large boulders and rocks, traversing steep cliffs and surface transitions, like a giant insect.  On smooth roads, it can "sit" on its 4 wheels, with its legs up in the air, drive and steer like a high speed 4WD vehicle at speeds up to 50 km/hour and even rotate on the spot clockwise or anticlockwise.  At present, the entire mechanical design has been completed and all robot control and simulation software has been designed and tested successfully, with 3D simulation movies to demonstrate.

AUTOMATIC OVERSPEEDING WARNING SYSTEM (2005):  I supervised a final year project that resulted the successful design, construction and testing of a radio-equipped speed sign that periodically transmits a maximum speed limit signal for a particular zone of traffic  (within a limited range of distance).  The “speed limit” signal can be detected by a radio receiver attached to a motor vehicle and can trigger an audible alarm if the computer detects that the vehicle is exceeding the posted speed limit being transmitted by the radio beacon on the speed sign.  Future research may focus on allowing such speed signs to “force” all motor vehicles (except Police, Fire, Ambulance and Emergency vehicles) to remain under the legal posted speed limit, within a particular zone of traffic.  (Videos of successful experiments are available).

SPI:  STRAYING PREVENTION INDICATOR (1996)   (patents filed) is a machine vision system that monitors the road lines near a motor vehicle, and triggers an audible alarm if a single line, double line or dashed line has crossed under the front wheels of the vehicle, indicating unintentional lane straying or roadway departure, one of the indications of driver fatigue, "sleeping at the wheel", or dangerous (careless) driving.  The SPI prototype was built and tested successfully for day and night driving conditions, and was demonstrated on Channel 9's "Today Show" several times in December 1996, in a 3 minute 40 second story presented by reporter Ms Karen Milliner.  The SPI device was endorsed by the Head of RACQ Insurance (Queensland), Mr Garry Fytes, and has the potential to save the lives of about 30% of all motorists who die on the road each year due to driver fatigue.  Unfortunately, none of the 20+ car manufacturers and electronic manufacturing companies that I approached showed any interest in supporting SPI commercialisation, but this invention appears to have been copied by a few major car makers (after I showed them my videos!), who have released similar performing products in just recent years, but using different and more expensive hardware.  My SPI prototype took about 2 years to develop and only costs about $200 in parts to build.  The first prototype took only 2 weeks to build, just to get the hardware and digital cameras working, but the biggest problem was removing all forms of false alarms being triggered by bad images and stray sunlight.  The prototype I have in my possession is now highly resistant to false alarms and can automatically adjust its sensitivity under all types of normal driving conditions using "software aperture" exposure techniques and 6 false alarm checks.  It is still possible to commercialise the SPI invention even today, but using newer components, since unfortunately, nobody in Australia was interested in funding this project for commercialisation, even though the SPI design has no disadvantages or problems.

AUTOMATED SOIL HARDNESS TESTING MACHINE (2005):  EVH Drilling Pty Ltd sponsored a masters project that I supervised, to design and build a new soil hardness testing machine which conforms to an Australian standard for lifting and dropping a fixed mass at a fixed height above a progressively descending anvil impact point, which pushes a tube into the ground.  The number of drops must be counted and the soil penetration must also be measured to 1 mm accuracy.  My sensor and controller design has been successfully implemented and built by my masters student and it had been tested successfully in very dusty environments.

AUTOMATED HARD HAT IMPACT TESTING MACHINE  (2002):  This was a final year project that I supervised for Mr Jonathan Fievez which resulted in the successful design, construction and commercialisation of a hard hat impact testing machine that conforms to Australian and NZ standard 1801:1997 (Appendix C & D).  Sponsored by R. R. Sales, Pty Ltd, Perth.

QTAR (2005) - Quad Thurst Aerial Robot:  This is a self-stabilizing 4-propeller flying UAV (Unmanned Aerial Vehicle) with VTOL (Vertical Take Off and Landing) capability.  I supervised the final year project of Joshua Portlock and Brett Hammill in 2005, which resulted in the successful construction and operation of an easy-to-control flying UAV, which can be remote controlled in 4 degrees of flight  (Altitude control hover:  up/down, Translation: forwards/backwards, Translation: left/right, and Yaw: rotation left/right).  The aircraft was fitted with an IMU (Inertial Management Unit) comprised of 2 gyroscopes and 3 accelerometers, which allowed the onboard AVR microcontroller to estimate pitch, roll and yaw angles and stabilize all flight manoeuvres.  The QTAR could carry an onboard wireless colour video camera to send back live video footage, and it can fly for around 15 to 20 minutes using onboard battery power, thus making it suitable for applications such as remote video surveillance and watching sports events like football games, close-up, from the air.

RARE (2005) - Remote Area Robotic Explorer:  This was a final-year project I supervised in 2005 which resulted in the successful design and operation of a wireless mobile robot (like a small army tank with 2 tracks) that could be remote controlled by a USB Joystick attached to a PC.  It was fitted with a front pivoting metal detector (inductive sensor) for detecting land mines, a wireless colour video camera that could rotate on a 360° controllable platform, and a Bluetooth wireless module.

OAS (ONLINE ASSESSMENT SYSTEM)  (2005):  OAS is a fully functional, secure online assignment editor and secure quiz system (which presents graphical assignments and automatically marks questions via the internet), demonstrating features that are similar to those of WebCT and Blackboard, but offering the potential to extend such features and custom design different methods of student answer inputs to best suit the subject matter.  (eg. Intuitive dragging and dropping of graphical objects using a mouse)  For example, lecturers can now set up pictures and assignment questions with randomized variable values unique for each particular student, student answers are entered via any web browser and checked using a correct equation, instant feedback is given on mistakes that they have made and their final marks are automatically recorded and can be printed in an end-of-semester summary report.  OAS can be extended to also allow password-protected on-demand video streaming and interaction with “live” 3D virtual-reality environments, characters and models of objects, similar to “Second Life”.

BLIND SEARCH INVERSE KINEMATICS METHOD  (1997, 2004):  This is a novel and highly robust (crash-proof) numerical algorithm that can always solve the joint angles/displacements (the IK solution) for any type of serial-link robotic arm or manipulator, regardless of the number of degrees of freedom, the types of neighboring links used (rotary, translational or spherical) or the kinds of internal singularities that are present within the manipulator’s workspace which would cause solution instabilities with almost all other IK solution methods.  3D Simulation tests have confirmed that this “Blind Search” algorithm is reliable, always stable and computationally efficient enough for fast real time control of any type of robotic arm.

WIRELESS SENSOR SUIT FOR SPORTS APPLICATIONS  (2007):  (Patented).  This is a new type of mobile, wireless sports suit that is covered with impact detecting sensors which can determine the location and intensity of impact shocks occuring over the user’s body, such as boxing glove impacts, karate kicks or high speed paintball impacts.  An onboard low-power microcontroller monitors all impact sensors and transmits GPS (Satellite Navigation) location data to a remote monitoring computer, so that a player’s exact location and the severity of simulated injuries can be ascertained, analysed and constantly monitored or observed by remote referees.  If the simulated injuries are excessive (ie. the paintball player, mixed martial artist, or boxer) has received too many “blows” or direct hits, and too frequently, the monitoring computer can command the player to leave the playing field or declare a “knockout” victory to the other opponent in order to avoid serious injuries.  Such a wireless sensor suit could be used in police, military, boxing and mixed-martial-arts training programs.  A working prototype had been successfully built and tested.  I first proposed this project and supervised 3 final year Mechatronics students to work on this suit since the beginning of 2007 and field trials were successful.

Examples of current and future research projects

Current projects:   Railcow 2 rail robot, Cowbot 2 mobile robot, EMLA 2 compliant linear actuator

Future projects:  FERMA compliant muscle, FETA robot arm, APE biped exoskeleton, CNC machine

Railcow 2  (for training “cutting” horses) - Current

This project was started in January 2009, and is currently fully designed and being manufactured right now.  This is a more powerful “commercial” version of John Billingsley’s first working prototype, built for horse trainer, Mr Mike Lawlor.  It will feature caliper disk brakes and a 300W motor which will deliver speeds in excess of 50 km/hour, to really challenge the most agile and powerful of horses.  All of its electronic components, motors and mechanical parts have already been purchased and the mobile platform has already been designed in AutoCAD.  Over 20 detailed and dimensioned machine component drawings were prepared and parts are currently being machined.  The entire Railcow 2 should be fully built and operational very soon.  The Railcow 2 platform may be used for other kinds of automation and robotics projects requiring linear motion control, such as automatic soil tilling, seed planting, irrigation and harvesting.

Cowbot 2: Mobile robot cow - Current

This project has not yet started, but the mechanical and controller designs are almost fully completed and most parts and components have already been purchased.  This is a general-purpose 4-wheeled remote controlled mobile robot with large 26” size mountain-bike wheels, powered by 2 x 750W DC electric motors and 2 x 12V car batteries.  Cowbot 2 will be able to drive at very high speeds forwards and backwards, stop instantly, rotate on the spot clockwise or anti-clockwise, and even steer left or right along a variable radius of curvature, like a conventional vehicle.  Cowbot 2 will have an onboard GPS (Sat-Nav) system to track its position and direction on an open field, and its Radio Controller will have a range of several kilometers.  Cowbot 2 can also be fitted with a wireless video camera for receiving streaming video images on a remote laptop PC.  It can also be programmed to “watch” the horse and rider (using machine vision), and move in the opposite direction that the horse is facing, just like a real cow.  All the main mechanical components, beams and materials for this robot have already been purchased, however, there is much CAD design, machining and fabrication work required to complete the idler gears and bearing housings for tensioning the drive chain.

 

EMLA 2:  Electro-Magnetic Linear Actuator - Current

This project was started in February 2009 by a final year USQ student.  I currently have the design for a working prototype along with demonstration videos, produced by former final year project students.   However, the very low force output for the first prototype design makes it only suitable for low force positioning applications.  (It currently acts like a weak pneumatic cylinder running at low pressure).   A USQ student final year student is currently developing and building a high-strength prototype EMLA which will hopefully be finished this year.  It is hoped that this new design will produce very high force outputs, and can be easily adapted to a flexible, bendable FERMA muscle design.

FERMA:  Flexible Electromagnetic Robot Muscle Actuator - Future  

This project has not yet started, but basic designs have been created.  This invention, to be called the “FERMA” (Flexible Electromagnetic Robot Muscle Actuator) will function and behave just like a human muscle, able to only contract or relax (release) with a variable force.  FERMA will be designed to be silent, lightweight, electric powered and be able to bend around corners, similar to an “air muscle” or “hydraulic muscle”, but without all the bulky valves, piping, reservoirs, manifolds, pumps or compressors.

It will be designed to enable robots involved in grinding and polishing operations to smooth out machining marks left over after facing, ball-nose cutting, pencilling or end-mill cutting operations performed by CNC machines.  At present, this kind of “finishing” work cannot be performed well by conventional CNC machines or typical 6-axis robot arms that use harmonic gear drives or large speed reduction ratios.  Such machines are typically too “solid”, too rigid  (ie. they cannot be back-driven easily) and they offer little or no “compliance” (or elasticity) in their joints to provide precise variable pressure / forces necessary for smoothing out surface undulations and wavy marks left over after roughing and finishing operations performed by a CNC machine or a conventional robot milling manipulator.  Variable compliance (or adjustable “springiness”) is needed to produce the best surface smoothing results, hence, well trained and experienced human workers (holding portable hand tools) are needed to perform such surface finishing operations.

 

FETA:  Flexible Elephant Trunk Arm - Future

This project has not yet formally started, but basic feasible designs have been created.  This project involves the design, 3D modelling, simulation, manufacturing, programming and testing of a FETA manipulator which will mimic the movements of an elephant’s trunk.  Several dozen FERMAs will be positioned around a hollow tubular proboscis manipulator (resembling an elephant’s trunk, or one leg of an octopus).  This flexible manipulator will be controlled by PC software to manipulate and control the 3D position, orientation and applied forces on a small rotary grinding tool, using custom-written inverse kinematics and muscle control software.

The FETA manipulator will demonstrate how FERMA technology can be used to perform high-quality “human-like” grinding and finishing operations on a large mechanical component that requires post-machining finishing operations.  The FETA manipulator will be very different to conventional rigid-link 6-axis robot actuators, because it will have the ability to weave and move around large obstacles, pass through narrow gaps and holes, reach almost any part of a casting or mechanical component from different approach directions, and achieve almost any end-effector orientation and position from a variety of different arm postures.  Such a flexible manipulator may be used for other practical applications, such as automated fruit picking or the harvesting of delicate vegetables or agricultural produce in an unstructured, dirty and dusty farming environment.  The FETA manipulator will also be able to wrap or twist itself around large objects, and grasp objects with an end pincher.  This will require a sophisticated control program to coordinate all FERMA movements and control the overall “shape” of the proboscis.

APE:  Adaptive Personal Exoskeleton - Future

This project has not yet formally started but basic designs have already been created.  With the help of postgraduate and final year students, I plan to design and build a giant humanoid biped robot suit that a person can “wear” which can increase human strength and endurance at least 3 times.  It will be entirely electrically powered using all-electric powered EMLA or FERMA muscles.  Also, it will be designed to be a stand-alone remote controlled biped robot that can be used for walking up and down stairs and performing tasks similar to the famous “Honda robot”.  Funding will be sought for building this project after good designs and 3D control simulations are developed.

CNC 5-axis Styrofoam sculpturing robot - Future

I will be working with Professor John Billingsley to prepare an ARC Linkage grant to fund the development of a 5-axis robot for slicing and sculpturing solid Styrofoam to serve as patterns for fibreglass moulding.  So far, I had consulted with John and shared good ideas about how to use MasterCAM or similar CAM software to automatically generate G-code tool-guidance text files automatically from 3D or surface models, which can be interpreted by Windows software to drive all the axes of such a sculpturing robot.  This is a fairly difficult task to complete, as it is equivalent to the task of designing and building a CNC milling machine.  It will be very labor intensive and take a lot of engineering and machining effort to build a high precision CNC “gantry robot” to control a high speed spinning blade to produce smooth surfaces along any 3D orientation or path in space.  This will be the subject of several PhD projects in the future.

If you would like to be involved in the commercialisation or development of any of the above current or future projects, please do not hesitate to contact me.  I am always eager to recruit and supervise enthusiastic and talented students who want to work on designing, manufacturing and controlling these kinds of practical projects.

My personal views about mechatronic engineering research in general