Minutes of Meeting 26th
May 2004.
Present: Jan Maciejowski,
Mark Mellors, Andrew Lamb.
Mark and Andrew: update on progress so far, particularly with frame design and chosen drive design.
Reported that now
approaching stage where building of frame can commence.
Brief discussion of budget breakdown. Totals about £280 so far for structure and mechanical drive system.
More detail to be
given to cost of electronic systems.
Range of budget to
be estimated, rather than specific amounts.
Identified sensors
as being a concern – more research is required.
Investigation of pendulum with linear accelerometer to give
change from vertical position – the previous work on this
dates back to 1985 and new technologies may be available.
Jan: job number given for
use with the project (specific information referring to
project required when used).
Andrew: No progress made on
PDA controller front due to dead-end responses from LCE
staff.
Jan: Book called PDA
Robotics was signed off – now on order with Pembroke
College library. Due to arrive in one week.
NOTE: Andrew has now bought
an HP iPAQ for himself! Should make the project a little
easier…..
Jan: Records must be kept of
all design decisions, calculations etc i.e.: why choose the
PDA route rather than FPGAs or PICs?.
Contact to be made with: Mick Furber (EIETL); Richard Taylor (Data logger project); Electronic Development Group (contact is Dr. Dukic); Dave Goutry (workshop near LR7); Oatley Lab (Andrew has contacts); Supervisor for 4th year project that used PDA controller system two years ago.
Next meeting proposed 2 weeks from today.
Last meeting of term proposed 2pm Thursday 17th May.
Meeting with Markus Kuhn
Micromechanical devices from Maxim or Analogue Devices, need ADC - Dynamic error?
External reference - laser range finders - time resolution - use accelerometers etc
Position in a room
Contact: USB device (serial interface) - email is to be forwarded to us
Markus would be interested in working on it!
Learning algorithm approach:
Discrete control
Interpolation/extrapolation from past experience
States: Angle, current speed
Brute force computing - searching
Piecewise linear
Catch outliers in prediction unit
Amount of experience needed?
Brute force method initially to produce a map of pointers for control
Fairly powerful computer needed initially
Intermediate goals on route and don't choose goals that are unreasonable
Two separate control systems in layered approach
State for environment (ground) parameter
Account for tyre pressure?
Robust! Programmer can't destroy electronics
Hang it up and control!
Glen Vinnicombe at
12noon in LT0 lecture!
Supervisions on
Mondays at 4pm
Webspace gives marks
for project organisation
18th to 25th Nov
range for talk
Schedule of tests for
'remote control wheel'
Speak to Mick Furber
about allocation of space for project
4F1 lectures may be relevant
Ball & Beam has
double loop feedback system
Risk Assessment
Warn Mr. Furber when using remote control
Block diagram of
system including double loop
Email MS Project plan to Dr.
Mac!
Storage space in
Part II EIST project lab
Bring the helicopter
report next week
Chase up Peter Long
for gyro specs
Speak to mechanics
lab about testing rate gyros
Check frequency of
IR sensor using carriage from inverted pendulum experiment
on sine wave
All velocity signals on controller rather than velocity and direction (to start with)
Dave Goultry (332764, dpg@eng)
to ask about dataloggers - knows about inverted pendulum kit
Do up Word systems diagram
on Simulink (sunday)
Get 'Bicycling Science' book
Dubious bit is conversion of torque demand into voltage demand: could have extra feedback of present acc.
(True if angular velocity is constant - might not matter is acc. is quick!)
Determine inertia of motor
& gearbox to test significance
Use simulink to figure out effect of other parameters
Flywheel speed from
tacometer on block diagram
Rolling road from
Allan McRobie
Monday 22nd talk at
12noon to 4pm
Fundraising letter
prep with budget
Interdependence of two control loops: leave open?
link from wheel speed to reduce gain of turn at high speed? Done?
Weight shifting using solenoids for fine tuning
Use mechanics lab
gyros to measure platform used in testing
Simulink help files on teaching system
Mathworks website examples: 'non-holonomic' - search!
Non-holonomic: need more variables to describe where it is than are available: forward, sideways, wheel turning
3-bar inverted pendulum (Japan) - 'look up Furuta'
check all gyros are
the same
256 levels of motor
control may be excessive
investigate 4QD
speed controllers
hardware interfaces to
reduce bandwidth
potentiometer to measure flywheel position
foam padding on panels may reduce noise/vibrations
Pembroke funding: some progress and ideas
but may take time. Matching funding best option
Next week's talk: Mark and I should not
contribute to each other's talks but can answer questions raised
Talk preparations: internet site
presentation; copies on USB, laptop and pen drive. 10 minutes talk time
Use hierarchical method for circuit system
diagram
Table for speed controller comparison -
Andrew
Speak to Patrick Palmer in Div. B about speed controllers and spares
The supervision this week was replaced by the project presentation
Speak to Malcom Smith about stability assessment
Rate Gyro??? Speak to Glen
Vinnicombe
Talk to Glen Vinnicombe
regarding choice of controller
Carmen Ho: Sim Mechanics
& Tyre friction models - set up a meeting
Research Kalman filter, bifurcation maps and poincaire maps
Could use Dave's image recognition project to track in X, Y
Enter delays into algorithm
loop on Simulink model
Film & Measure angle of lean?
Compile Simulink to C++??? Or
just C? Ask John Else
Investigate Sharp and Futaba for funding
RAEng and IEE grants?
Try using a smaller motor in the simulink model
Add friction & backlash into the model
Get in touch with Hugh Hunt's student
Change model to make angles relative to vertical
Investigate dynamics of rigid bodies w.r.t. angle transformations
Focus point for thought is Euler angles
Speak to Dave Miller about using the
mechanical design workshop space
Put Gyro results on
website
HTML version of our reports on website?
Longer gyro test for drift to answer whether effect is a time delay
Correct the equation that was given incorrectly in Mark's report
Get Mark to email
latest photo
Order Active Robots
through EIETL and new job. no.
CANCELLED
National Instruments Demonstration
Thursday 10th February
Business card for Ian Matthews
Can arrange training with Ian for Wednesday 16th Feb
Templates are available form 'new' button
Front panel is basic interface
MAX can be used for connection and has to detect FPGA
Need to configure the FPGA and processor interface
Can run wizards to split into real time, Windows, comms etc
'Project' approach rather than hardware approach
Q: PDA module available?
Q: Is the 'three' FPGA the 'big' FPGA?
Can have home licenses from Uni license so can install on laptop
Extra module for importing from Simulink
Speak to Jorge
Goncalvez about borrowing extra analogue input module
Speak to Dr. Britto
about LabView training CDs
Get dimensions of
CompactRIO
Order Active Robots
myself and claim back
Postponed
Postponed
Met in workshop to go through progress
Ask Computer Operators for extra RAM
Away April 10th to start of term
Deadline for balancing forward/back is 10th April
Phone NI Technical Support!
Help System for signal conditioning issues