Minos 2018 was held over the weekend of April 7/8 at the Quality Hotel, Coventry. New venue, several new faces but the expected collection of great presentations and activities for micromouse and small robot builders.
The schedule ran like this:
09:30 Arrival. Coffee/tea
09:50 Duncan “Administrivia”
10:00 Peter Harrison “Competitions round-up”
Peter Harrison presented a range of photos gathered from recent contests in Taiwan, Japan and the United States highlighting some interesting robots and events. These are a regular feature of Minos and serve to help keep everyone up to date with the latest developments from around the world. The gallery can be seen here: https://micromouseonline.com/2018/04/10/minos-2018-roundup-slideshow/
10:45 Ian Butterworth and David Hannaford “Registration and Timing System(RATS).
Ian and David have been working hard on a new, portable timing system for micromouse and related contests. David has worked on the gate hardware which is linked by radio to the software written by Ian. Novel features in the gate hardware include a presence detection in the start square that makes it possible to automatically sequence the timing system without the need for user intervention for re-arming and run detection. The software is written to permit the registration and timing for a variety of events and the incorporation of event-specific scoring rules. A defined serial communication protocol make the system flexible as well as encouraging the production of alternate timing hardware solution that can operate reliably with the recording software.
MINOS18 David Hannaford
MINOS18 Ian Butterworth
11:45 David Otten “Line Follower”
David Otten’s Jehu line follower has been getting some more attention lately with upgrades to the sensor designed to reduce the centre of mass. Jehu’s sensor is unique in line followers in that it uses a sing Position Sensitive Device from Hammamatsu, and a mirror to allow it to see the line with a single sensing device and give an accurate indication of error from just two ADC measurements. In his presentation, David describes the latest improvements and explains how the sensor is used and calibrated.
MINOS18 David Otten
12:15 Alex Bagehot “Impressions of a new mouser”
Alex is a newcomer to micromouse. Today he gave us his impressions of the community and an outline of what he hopes to achieve as he gets more involved in building his own robot.
12:35 LUNCH bring a packed lunch or find something locally.
13:30 Duncan Louttit “Maze Making”
To build and test a micromouse, you will need a maze. Full size mazes are 3m square and not easy to fit into the average house. A modular maze design allows smaller practice mazes to be combined into varying sizes including a full-size classic maze suitable for contests. By creating the maze from nine identical tiles, Duncan has made it easy for a newcomer to build as much maze as they can handle. Careful construction and material selection, along with adherence to a few simple dimensional constraints makes it possible for pairs or groups of builders to keep their own sections and then get together at events, joining their individual parts together with confidence that they will all fit together properly.
This modular maze system is designed to be relatively inexpensive and easy to transport in most family sized cars. The whole maze could easily be moved and set up by one person if necessary.
The presentation included images of the construction and a full guide to making your own.
13:50 Michael Beatus “My next Micromouse especially the sensors”
After a bit of a break from micromouse robot building, Michaels is keen to get started again. He shared with us some of his thoughts about recent developments and the influence they are likely to have on his immediate plans for a new mouse.
14:20 Bernard Grabowski “Orange Mouse”
the “Orange Mouse” from Bernard is a multi-purpose robot design intended to be suitable for a range of small robot contests. Simple, robust design with optional sensor components makes for an ideal introduction for schools and beginners. With inexpensive motors and other components, #D printed parts should mean that almost anyone can have a go at making one for contests. They don’t have to be orange – Bernard happened to have a lot of orange filament to hand so he used it. Perhaps he has his eye on 2020.
MINOS18 Bernard Grabowski
14:35 Derek Hall “Sensor calibration”
The micromouse, like any robot, is very dependent on the sensors for correct operation in the maze. While micromouse sensors are simple in design and operation, the results need careful interpretation for reliable outcomes. Sensor calibration and characterisation can be a tedious and error prone business so Derek Hall and Jim Chidley decided to introduce some automation to the process. They took a few commonly available parts like stepper motors, leadscrews and the inevitable Arduino and coupled the lot to an Excel spreadsheet. Their calibration rig can automatically gather sensor readings as a function of distance and plot the results in a spreadsheet for further analysis.
MINOS18 Derek Hall Sensor Calibration
14:55 Alan Dibley “Large scale 3D printing”
Most people who want to get stuff done with 3D printing are happy to go out and buy one and get printing. Then again, most people do not want to print parts as big as Alan Dibley. Literally in fact – Alan’s 3D printer could make a full size replica of his own body if he wanted to! He already had a large home-made gantry mill so it seemed like a natural extension to add an extruder and turn it into a 3D printer. While he was at it, he also write much of the software from scratch since existing code was not really suited to the main purpose he had in mind which was printing wings for slope soarers. In his talk, Alan described the evolution of the printer and the unique software that drives it. Here are a few photos.
15:10 Peter Harrison “Edumouse: the Trainer from Taiwan”
Building an affordable micromouse that also has good performance is quite a challenge. Many of the top robots use very expensive motors with built-in high resolution encoders. These are not a realistic choice for the beginner and certainly not suited to a training mouse intended to be build in larger quantities for student use. In Taiwan, the team at Lunghua University have designed a micromouse that uses very cheap motors and low resolution encoders. These encoders are combined with information from an accelerometer to provide an very good estimate of the robot position that allows for smooth control at high speeds and with good precision. The technique used – a state space observer – is novel and much simpler to understand and implement than the Kalman filters that many believe to be essential. Experimental results demonstrate that there is no significant advantage to the use of a Kalman filter in this application. Anyone could build a better robot with the observer-based estimator.
In his presentation, Peter Harrison describes some options for implementing low resolution encoder systems and explains how to make the most out of the observer-based estimator and feedforward to achieve very good motion profiling in a robot that, altogether, costs much less than one of the motors in a typical top-flight micromouse.
MINOS18 Peter Harrison
You can see Edumouse run here: https://youtu.be/pg3-_mP_jJU
15:40 Stephen Pithouse show and tell
Stephen has been working on small robots with great success since he was at school. Now he has put some of that experience into practice and is working for FANUC who claim to offer the widest range of industrial robots in the world. In his talk, Stephen gave us some insight into work at the FANUC factory.
15:50 Garry Bulmer “ESP32”
In recent years, there has been a huge growth in wireless communications. From basic WiFi, the technology has expanded to the point where nobody is even surprised when you tell them that all the lightbulbs in your house can be individually controlled through the house wireless network. As an offshoot of this huge growth there are many devices now available for very little money that make it possible for the curious to incorporate sophisticated wireless communication into almost any kind of hobby-level project. Many of these were based on the ESP8266 chipset. This tiny marvel turns out to be a fully programmable microcontroller that can be programmed using the arduino environment. In case that was not enough, the originators of this chip have brought out the ESP32. This is a 32 bit, dual core powerhouse with just about every peripheral you could think of already built in as well as, of course, a full set of wireless protocols and hardware. Garry Bulmer is clearly excited by the possibilities such a device can bring and explained to us many of the available features.
Soon there will be slides!
16:10 Powerpoints from Balazs Gonter and Ian from Pembrokeshire Robot Builders Club who are unable to attend.
Difficulties with travel from Hungary meant that two of our overseas contributors could not make it to the actual event. We hope to see them very soon because they have clearly been hard at work with a great deal of experience to share. They did, however, send in their slides and we did what we could to interpret their content. It was in English but we had to work out some of the content for ourselves. As it turned out, that was an interesting exercise that turned up many questions and promoted some debate. This is what they sent:
MINOS18 Halfsize Micromouse from Hungary
Pembrokeshire Robot Builders Club were also unable to attend but sent in some slides about their Mechanical at robot project. It looks pretty cool. You can find out more directly at mechanicalcat.org
MINOS18 Mechanical Cat
Many thanks to Duncan Louttit for all his efforts this year in organising a most enjoyable and successful Minos conference as well as to Ian Butterworth for carting the large BCU maze about for the very last time.
Sunday Contest Results
On the Sunday morning, we ran a small informal contest. The results can be found in this file (thanks to the timing system from David and Ian): Minos 2018 Results