From 7b76b2c73d53a2ed31e3c887926f44dd02de44d2 Mon Sep 17 00:00:00 2001 From: lonkaars Date: Wed, 10 May 2023 15:20:35 +0200 Subject: added research --- doc/base.tex | 6 ++-- doc/research.md | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 2 files changed, 96 insertions(+), 4 deletions(-) diff --git a/doc/base.tex b/doc/base.tex index ecc3c8d..9c1c908 100644 --- a/doc/base.tex +++ b/doc/base.tex @@ -6,7 +6,6 @@ \usepackage{siunitx} \usepackage{amsmath} \usepackage{csquotes} -\usepackage{parskip} \usepackage{unicode-math} \usepackage{fontspec} \usepackage{tabularx} @@ -27,7 +26,10 @@ \bigskipamount=7mm \medskipamount=4mm -\parindent=0mm +% \parindent=0mm +\parskip=1ex + +\def\tightlist{} \title{\doctitle} \author{ diff --git a/doc/research.md b/doc/research.md index 1a616f9..ba697a8 100644 --- a/doc/research.md +++ b/doc/research.md @@ -1,4 +1,94 @@ -# test 1 +# Problem statement -this is a test document in markdown that will get converted into latex +The following is the original project description (translated to English): +> I would like to bring to market a vehicle that can drive independently from A +> to B. The vehicle must take into account traffic rules, road signs, traffic +> lights, etc. Research is being conducted using a small cart, the Pololu Zumo +> 32U4, on which a camera module Nicla Vision is mounted. The aim is to +> investigate the most appropriate method of recognizing the road, traffic +> signs and traffic lights. This should be demonstrated with a proof of +> concept. The cart does not need to drive fast, so the image processing does +> not need to be very fast. Assume one frame per second (or faster). + +# TBD: The big question + +## Communication between the Nicla and Zumo + +In order to make the Zumo robot both detect where it is on a road, and steer to +keep driving on said road, some sort of communication needs to exist between +the Nicla and Zumo. As mentioned earlier\footnote{dit is nog niet benoemd}, all +machine vision-related tasks will happen on the Nicla board. Because the Nicla +board is the first to know how much to steer the cart, it makes sense to have +it control the cart by giving the Nicla a 'steering wheel' of sorts. + +This section tries to answer the question "What is the best protocol to use +over the existing UART connection between the Nicla and Zumo?". After a +brainstorm session, we came up with the following specifications for the +communication protocol: + +1. **Low bandwidth** + Less data means more responsive steering +2. **As simple as possible** + The Nicla only needs to control speed and steering +3. **Easy to mock and test** + The cart should be able to be controlled using a mock driver and the Nicla's + output should be testable (preferably using unit tests) +4. **Adaptive to noisy data** + The cart should gradually change speed and steering direction as to not slip + or cause excessive motion blur for the camera module on the Nicla +5. **Adaptive to Nicla failure** + If the Nicla crashes or can't detect anything, it will stop sending control + commands. In this case, the Zumo robot should slowly come to a halt. + +Where possible, it's generally benificial to re-use existing code to save on +time. Existing code exists for a custom binary protocol and a text-based +command protocol. Both of these were designed without bandwidth or latency in +mind, and mostly focus on robustness in the case of temporary disconnects or +noise on the communication lines, so a new protocol needs to be made. + +To address specification 1 and 2, the command length is fixed at 1 byte. This +means that UARTs built-in start/stop bit will take care of message start/end +detection, since most software interfaces for UART (including Arduino) string +multiple sequential messages together even if they're not part of the same UART +packet. + +To mock messages from the Nicla to the Zumo robot, a simple USB serial to UART +cable can be used, along with a small C or Python program to convert +keyboard/mouse input into steering/speed commands. A small software layer can +be implemented on the Nicla to log the semantic meaning of the commands instead +of sending actual UART data when run in a unit test. + +A PID controller can be used to smoothly interpolate between input +speed/steering values. This would also introduce some lag between when the +Nicla knows how much to steer, and when the Zumo actually steered the wanted +amount. Smoothing the speed/steering values does make it virtually impossible +for the Nicla to make it's own input data unusable because of motion blur, so +the lag needs to be handled in some other way as directly controlling speed +values without interpolation would lead to a garbage-in-garbage-out system. The +simplest solution to motion blur is limiting the maximum speed the Zumo robot +can drive at, which is the solution we're going to use as speed is not one of +the criteria of the complete system\footnote{Problem statement +(\ref{problem-statement})}. + +In the case the Nicla module crashes or fails to detect the road or roadsigns, +it will stop sending commands. If the Zumo robot would naively continue at it's +current speed, it could drive itself into nearby walls, shoes, pets, etc. To +make sure the robot doesn't get 'lost', it needs to slow down once it hasn't +received commands for some time. As mentioned in section \ref{TODO}, the Nicla +module is able to process at about 10 frames per second, so 2 seconds is a +reasonable time-out period. + +\def\communicationConclusion{ +The complete protocol will consist of single byte commands. A byte can either +change the cart speed or steering direction, both will apply gradually. When no +commands have been received for more than 2 seconds, the Zumo robot will +gradually slow down until it is stopped. Exact specifications of commands are +provided in the protocol specification document\footnote{dit document bestaat +nog niet}. +} +\communicationConclusion + +## Conclusion + +\communicationConclusion -- cgit v1.2.3