Build A Mobile Robot from Scratch —— Part I

1. Intro

Why this project? I wanna build a robot from scratch to fill the gap between theory and reality.

To build the robot, usually need following knowledge base

Mechanics, the build the body of robot
Electronic, the power of the robot
Computer, the brain of the robot

2. Goal

For the robot, I would like to focus on the mobile robot platform

Two level goals:

Basic level, the robot can move, and use the lidar as a SLAM sensor, can track something using camera
More precise control, going from A to B, the error should be less than 1cm

Consider a layer platform

Environment sensing, different sensor to be used, a SBC is used. LiDAR, video camera, speaker, microphone
Intelligence, communication, SBC, control other part
Navigation, get the high-level command and go to desired position

Communication Ethernet

3. Path

Motor

Control how it works
feedback control
position control
speed control

4. Chassis

Design

34cm 30cm 23cm

Front x2

Rear x2

5. New Thing

5.1. Motor Driver

Ref

Working with DC motor

Direction
Using the H-bridge could change the direction of voltage and current
MOSFET would drop a little voltage
Speed
PWM

L298N, simple, classic but too old

More improved design:

TB6612FNG, improved version of L298N
DRV8871, if you only need a SINGLE motor to drive
MX1508, small, but not soldering friendly

Driver that can handle large current:

DBH-12，20A, peak 30A
IBT2, single channel, 30A, peak 43A

5.2. Micro Controller

UNO or raspberry pi PICO

Pico should be a better choice in 2022 (introduced in 2021)

Much faster (133MHz vs 16MHz)
Much larger RAM
Lower power DC (1.8v-5.5v)
More port (Analog and Digital)

5.3. Battery

LIPO battery, high capacity in lightweight

Con: easily damaged

Lithium-Ion Polymer (锂离子聚合物电池)

Application where weight matters

Each cell can provide 3.7V

Fully charged 4.4V
Minimum safe charge is 3.0V

Different wire

S- serial
P- Parallel

Capacity

C rating

Connection:

Dean
XT60
EC3

5.4. Base Material

Consider: strength to weight

Aluminum

lightweight

Acrylic (亚克力)

lightweight
drilled

5.5. Channeling System

Actobotics system: good availablity

GoBilda system: good quality, successor to the Actobotics

Vex system

Lynxmotion

5.6. Motor Selection

DC Gear motor

Planetary: a bit expensive, but quiet
Spur: cheap, but noise

Things to consider

Nominal voltage, optimal voltage
No Load RPM, when nothing connected, what is the RPM
Stall torque and current. Newton Meter
Stall: no more move
<30% stall point
Size and shaft
Mount
Shaft may not in the center

Encoder is important

We need to first determine the

Weight (Load)
Speed
Longest working area for the robot
The robot should be slower than human walking (safe to control the robot)

Calculate the required specification for the motor and battery

Calculator

Motor, hub, bearing, encoder

Bearing is quite important since the whole load is on the wheel
if there are no bearings, the motor would load all the pressure

5.7. Wheel

Parameter for wheel:

Wheel size, determine the speed
Wheel width, determine the traction and maximum load
Wheel traction, determine the surface robot can travel
Wheel type, determine the degree of movement

Type

Standard, most common, wide range size, can be used as driver or idler
two degree (forward and backward)
Caster and Ball wheel, orientable, use as idler, most used for indoor design
Mecanum, different left and right wheels, multi-drirectional, 45 degree roller
Omni wheels, 90 degree roller, 4 degree of freedom, best for indoor

Arrangement

Durometer

harder or softer？
30A maybe too soft

5.8. Electronics

Motor driving

Power distribution board

Motor controller

Fuses: (保险丝)

if the current is too big, would case heat, may lead fire

5.9. Power Source

Ref

Regulator and Convertor

Logic voltage, 5V(4.7-5.2), 3.3V.

Motor: 6V, 12V

AC:

DC:

From high to low:
voltage regulation，AMS1117-5(much lower voltage drop), 1A
buck converter, high-efficient, mini360
From low to high
boost converter, PSM205, to a usb
From any to any
Buck boost converter, S9V11F5 (from 2-16 to 5)

5.10. Motor Controller

5.11. Weight Consideration

If the robot is too lightweight, there would be not enough traction for the robot to move.

5.12. Connector

JST

6. BOM

Part	Name	Num	Check
Front wheel	Motor	2	1
Front wheel	120mm Wheel	2	1
Front wheel	8mm REX bearing	2	1
Front wheel	8mm Face bearing	2	1
Front wheel	8mm shim	4	1
Front wheel	m4 washer	8	1
Front wheel	Hyper Hub	2	1
Front wheel	12mm M4 screw	8	1
Front wheel	18mm M4 screw	8	1

6.1. Omni wheel

Name	Num	Check
Omni wheels	4	1
Spacer 4mm	2	1
6mm D shaft hub	2	1
6mm D shaft bearing	4	1
6mm Collar	2	1
30mm M4 screw	8	1
80mm 6mm D-shaft	2	1

7. Progress

11.21

12.5 Test Robot Motor

Full Speed
Arduino PWM
20%
50%
100%
Encoder
Each rotation

TODO:

Add a ferrule crimping tool kit

Motor Driver:

VCC is connected to 5V power supply of MCU,GND is connected to GND of MCU.

R_EN AND L_EN short circuit and input PWM signal for speed regulation

L_PWM,pin input 5V level,motor forward

R_PWM,pin input 5V level,motor reversal

void setup() {
// put your setup code here, to run once:
pinMode(8,OUTPUT);
pinMode(7,OUTPUT);
digitalWrite(8, HIGH);
digitalWrite(7, HIGH);

//PWM output
pinMode(6,OUTPUT);
pinMode(5,OUTPUT);

}

void loop() {
// put your main code here, to run repeatedly:
analogWrite(5, 50);
delay(1000);
analogWrite(5, 100);
delay(1000);
analogWrite(5, 150);
delay(1000);
analogWrite(5, 200);
delay(1000);
analogWrite(5, 255);
delay(1000);
analogWrite(5, 0);
delay(1000);
analogWrite(6, 50);
delay(1000);
analogWrite(6, 100);
delay(1000);
analogWrite(6, 150);
delay(1000);
analogWrite(6, 200);
delay(1000);
analogWrite(6, 255);
delay(1000);
analogWrite(6, 0);
delay(1000);
}

7.1. Update Dec.10

Currently the robot works great on open-loop rotate and move forward, backward

Close-loop control on the RPM

Test how linear the motor is. That is the relation between the PWM duty rate and the RPM

7.2. Update Dec.13

Re-organize the wire and hide the wire in the alumni channel

Currently the robot’s speed control is under PID control

To make the position control closed-loop, still need encoder+IMU

And localization system should be added (SLAM)

7.3. Update Dec.15

Use DC jack to power the Jetson and Arduino

Given the hardware and electronic elements I have, followed step

~~Arduino Remote Control~~
~~Jetson use~~
~~Communicate with computer~~
~~Setup wifi enable remote access~~
~~Read IMU data~~
~~Install Lidar and read lidar data~~

7.4. Update Dec.17

~~Rewire the arduino~~
~~Power Arduino with Jetson Nano~~
~~Communicate between arduino and jetson nano~~

7.5. Update Dec.19

~~Add IR remote in arduino~~

TODO:

~~Remote control move forward/reverse/rotate~~
Use the lidar to build the map

7.6. Update Dec.21

Remote speed control and stop done

TODO:

Add IMU
~~Wire~~
~~Read data~~
~~Calibrate~~
~~Store offset~~
~~Reload offset~~
3D print a level for lidar
SLAM

8. Takeout

Low frequency PWM would cause the motor hauling, by increasing the frequency (like 3k) the hauling would disappear. In some MCU, by playing with the register, we might change the default frequency.

While selecting the MCU, checkout how many interrupt are needed and how many interrupt can the MCU provide.

By using encoder’s both channel and both rising and falling, we can maximize the accuracy of the motor

A good motor is almost linear between RPM and PWM

When the code doesn’t go well, check the setting and setup

After rewiring, double check the position especially the pin number

While making the wire, the length or wire should be considered. It’s a system design, so leave some room for wiring but not too much