Some thoughts on additions to ArloBot

[thomasco...@gmail.com]

Hi Chris,

Have you ever thought about adding a mechanical arm to your ArloBot like the one on the Heath Hero 2000.

Also, I think that it might be easy to set ArloBot up to seek out a charging station when the SLA batteries get low.

I would be willing to work on these two projects with you if you are interested.

I can do the mechanical/electrical side while you provide the programming.

Comments?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Here are a couple of links to examples of robot autodocking with most using ROS:

Link 1

Link 2

Link 3

Link 4 Somewhat dated, but good description of theory, design, and implementation.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

This project guide in this link might be the simplest and easiest autodocking scheme to implement. 

It uses ROS navigation to get to the "landing point" and then is guided to the dock by output from two ir sensors, separated by barrier, looking at an ir emitter on the docking station.

An Arduino Uno R3 is used to modulate the ir emitter and an Arduino Uno R3 is used to detect the outputs of the two ir receivers and provide control signals to ROS.

Comments?

Regards,

TCIII 

[thomasco...@gmail.com]

Hi Chris,

Some additional thoughts about the project link above: 

The RC5 transmission protocol is designed for 36kHz and it might be best to switch to the NEC protocol which is designed for 38kHz.

IR receivers and led IR emitters for 38kHz are much more common.

It will mean rewriting the Arduino IR receiver code for the NEC protocol instead of the RC5.

Comments?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I have received these IR Emitter/Receiver sets and have a couple of Adafruit Metro 328s that I can use to test out the IR emitter and IR receiver hardware and code this week.

Do you think that you might have time to take a look at the autodocking ROS code and see if it can be integrated into the ArloBot/ROS software package?

If this autodocking project looks doable, I can build up a docking station that uses a Power-Sonic SLA 12 volt battery charger and determine what it will take to design and build a mechanical charging coupler between the ArloBot and the charging station.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

TCIII

[Christen Lofland]

This thread seems to have a lot of stuff in it. My thoughts quickly:

1. Auto-docking:

The primary issues I see is how to safely make the electrical connection. In the past my robot was plugged into a 110v AC Outlet in order to charge the battery, but now that it is just the Pi, I only need a 12v DC charger. That seems much safer and easier to work with, but we still need some physical design for a "dockable" 12 volt electrical plug.

Unplugging is easy, just make sure the plug is angled correctly and drive away.

So if you have ideas of how to make a 12 volt DC connection that could be "driven onto" or "into", that would be the start.
Ideally something that could be relatively easily adapted to multiple platform shapes and sizes.

Some sort of 3D printable system would be nice.

Just for run and reference, ROS was originally created by "Willow Garage" and the robot they built for it was the "PR2", which actually plugged itself into the wall!
https://www.youtube.com/watch?v=GWcepdggXsU

Unfortunately, the last I checked that robot costs $250,000 and is also discontinued.

The image in the link you sent: https://github.com/wennycooper/Auto-Docking-Design looks like a similar design to the robot vacuum I own, it uses IR to line itself up to drive onto the base. The vacuum and base though are highly engineered to work together once the vacuum drives onto it.

Finally, no, I haven't looked at any of the ROS docking code or options. I'm willing to if we have some valid dock-able hardware ideas.

2. Arms

Arms are an entirely other thing. There is a ROS subset of tools called Moveit dedicated to arms and the like: https://moveit.ros.org/

I have never dug into it, as it felt like just building the arm was going to be more effort than fun. That said, I also didn't have  3D printer back then.

I'm vaguely interested in the topic, but there is a lot of planning and design involved to even get started on that.

[thomasco...@gmail.com]

Hi Chris,

Thanks for the detailed response, much appreciated.

As for setting up ArloBot for autodocking, I think that it is best to start with getting ArloBot to dock successfully with a base and then tackle the battery charging part.

The battery charging part of autodocking is pretty straight forward, but not simple unfortunately. The SBC cannot be running while the battery is charging therefore requiring a way to shut the SBC down before charging begins and then start it up after charging is complete.

I am thinking of providing around 14 vdc from the charging dock to the robot and then have a microcontroller shutdown the SBC, charge the battery, and when charging is complete restart the SBC. 

I would appreciate it if you could look at the autodocking hardware/software concepts in this autodocking project as the hardware concept seems solid

and the ROS software seems like it might be an easy integration into the ArloBot software package.

Comments?

Regards,

TCIII 

[Christen Lofland]

Personally, I do run the SBC while charging. My goal is that the SBC is always running, because that is where I can send/receive telemetry about the robot and allow for decision making. I'm not sure why we couldn't have the SBC running? Unless maybe your charger isn't putting out enough amps to both keep the SBC running and increase the battery's charge? I know that can be an issue, but I think that the solution is to get a charger that puts out more amps, up to a point. That or maybe your SBC is very power hungry?

Can you post the link again? When I click on it don't get to a working page.

[thomasco...@gmail.com]

Hi Chris,

Thanks for the quick response, much.

Here is the link again, sorry about that.

Leaving the SBC running can be done; it just means providing a higher current 16 volt source to power both the SBC and the battery charging circuit at the same time.

All of my 5 vdc switching regulators that power the SBC and the USB 2.0 HUB can handle up to 24 vdc. 

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I was able to build an IR transmitter using an Arduino Pro Mini that continuously transmits a Sony "power on" code of 0xa90.

I could have just had the IR transmitter transmit a continuous 38kHz signal, but that might be prone to interference from external IR sources and the "power on" code guarantees that the IR receiver sees only the IR transmitter.

Comments?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I have started on the IR receiver assembly which consists of two IR receivers and an Adafruit Metro 328 Arduino clone topped with a prototyping board. 

The prototyping board has two three pin connectors that supply power, ground, and signal connections from the prototyping board to the two 38kHz IR receivers that are mounted on a piece of perf board and separated by a divider shield.

The perf board will also serve as a structural mount for the two IR receivers. The divider shield is used to isolate the two IR detectors from each other so that the IR receiver code can determine when the robot is aligned to the middle of the IR transmitter which is centered in the middle of the docking station.

I have modified the existing Wenny Cooper's Arduino IR receiver code to eliminate the ROS commands for the purpose of testing the IR receiver performance with the IR transmitter. 

Once I am sure the IR receiver performs to expectations, I will revert to the original "IR_receiver_rosserial_ircode.ino".

Comments?

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I am getting ready today to do a manual (not on robot) test of the docking station IR transmitter/receiver system to see how it performs.

If all goes to plan, I will begin construction of the docking station to house the IR transmitter and the charging contacts. Found a good analysis of the design and testing of different types of charging contacts here.

One thing I noticed is that Wenny Cooper (Auto-Docking-Design) expected to use a limit switch to determine when the robot was securely in place on the docking station after acquiring the IR transmitter beacon.

However, limit switch detection is not present in his "mybot_autodocking_irCode.cpp" code as the function is put to "sleep" once the docking station is found.

So it looks like we will need to add some ROS code that detects the state of the limit switch and proceeds to initiate the charging function?

Comments?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Okay, I got the prototype IR transmitter/receiver assembly working to expectations in that the IR receiver is correctly decoding the IR transmitter modulation of the 38kHz IR beam.

Now to mount the prototype assembly at the same level as they would be on the robot and the docking station and experiment with the docking station distance and angular offset of the robot to the docking station.

Comments?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

If you decide to use and modify Wenny Cooper's Arduino code like I did for the IR transmitter and receiver, then do the following:

1) Because his code is so old when you use the Library Manager to install 'IRremote in the Arduino Library, make sure you install v 1.0

2) Then download Arduino-IRremote from here and copy and paste 'CppList.h' into the IRremote library in the Arduino directory.

3) I used the attached IR Transmit file to transmit the Sony 'power on' code and then removed the ROS references, and changed 0x43 to 0xa92, and uncommented the Serial println commands in Wenny Cooper's 'IR_receiver_rosserial_ircode.ino' code so I could view the output of the two IR receivers.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Wenny Cooper's IR receiver code turned out to be unreliable in being able to decode a particular IR code such as the Sony "Power On" code that I was trying to use.

After reviewing Chris Schurs' Robot Docking Logic project, I determined that I could reliably just use a continuous 38kHz beam without any kind of code modulation.

So I setup my IR transmitter to just transmit a 38kHz continuous carrier without modulation.

Unfortunately the TSOP38238 38kHz IR receiver that I am using does not like a continuous carrier and will turn off its output after about a second when it detects a continuous carrier without modulation.

Fortunately for me other users of the TSOP family of IR receiver have experienced the same issue and have come up with various solutions one of which I have successfully employed for this docking project.

I am now able to point the LED IR transmitter at the dual receiver assembly and get 'True' outputs from both detectors when the IR beam is being projected directly between the two IR receivers.

Like wise, moving the IR transmitter beam to either the right or the left results in a 'True' for the IR receiver seeing the IR beam and a 'False' for the IR receiver not seeing the beam.

The two detectors are separated by a divider vane and the IR transmitter has it 34 degree transmission beam spread reduced by a tube of shrink sleeving.

Therefore I have been able to duplicate Wenny Cooper's "IR_receiver_rosserial_ircode.ino" code using just an unmodulated 38kHz IR beam.

So how about you taking a look at Wenny Cooper's "mybot_autodocking_irCode.cpp" ROS code to see if it can be integrated into your ArloBot/ROS software and provide an autodocking feature?

And yes, I am working on identifying the best and safest way to transfer the battery charging voltage between the docking station and the robot chassis.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

Comments?

Regards,

TCIII

[Christen Lofland]

I will try to take a look at Wenny Cooper's "mybot_autodocking_irCode.cpp" ROS code this week.

Fortunately, part of the purpose of ROS is to make it relatively simple to integrate multiple inputs like this.

[thomasco...@gmail.com]

Hi Chris,

Thanks for the response, much appreciated.

I originally picked Cooper's Autodocking Design because of the simple IR interface and the ROS node software.

However, the IR transmitter/receiver design left much to be be desired and required extensive rework as my previous posts indicate.

Hopefully the ROS Autodocking node that Cooper wrote will easily integrate into your ArloBot ROS software with a minimum of effort.

Issues that still need to be addressed:

1. I assume that the robot will approach the dock head-on from the front and not attempt to back in like the Roomba/Shark does as this affects where the IR Receiver is mounted in the robot chassis?
2. The type and location of the charging contacts. See this autodocking analysis.
3. Providing power to both the battery charging circuit and the robot SBC/Aux electronic hardware simultaneously.

Comments?

Regards,

TCIII

[Thomas Coyle]

Hi Chris,

Find attached three pictures of my IR emitter/receiver project.

The first picture is of the IR receiver head that shows the two IR receivers, two LEDs that indicate when each IR receiver is receiving the IR transmitter 38kHz carrier, and the divider vane.

The second picture is of the IR emitter (transmitter) beacon. The IR LED is covered by an inch of shrink sleeving that helps to narrow its wide beam width of 34 degrees.

The third picture is of the Adafruit Metro/Prototype HAT assembly. The IR receiver head is connected to the Metro/Prototype HAT through two three wire cables. The resistors/capacitors on the board form a lowpass filter for the the output of the two IR receivers.

Regards,

TCIII

--
You received this message because you are subscribed to a topic in the Google Groups "ROS for Arlobot" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ros-for-arlobot/UbCetuR-h5A/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ros-for-arlob...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ros-for-arlobot/df350e46-1e24-44f0-be25-9b09669f6208n%40googlegroups.com.

[Thomas Coyle]

Hi Chris,

Find attached the Arduino C code for my IR transmitter and receiver. The IR Receiver code needs to have the ROS commands added back in per Cooper's original "IR_receiver_rosserial_ircode.ino" while the IR transmitter code just needs some cleanup to remove commented code that is no longer used. 

Also find attached the schematic for the IR transmitter (TSAL6200). 

1) The series resistor between +5 vdc and the anode of the IR LED depends on the forward voltage drop of the IR LED and the desired LED current flow.

2) The switching transistor is a 2N2222A/2N3904 and the base resistor is 4.7k.

The IR receiver (TSOP38238) schematic is the same as Cooper's with the addition of the following:

1) A 0.1 mfd ceramic capacitor across the IR receiver's GND and Vs.

2) A 1k resistor in series between the IR receiver output and the input of the Arduino and a 10uf tantalum cap from the Arduino side of the 1k resistor to ground which acts as a low pass filter.

Regards,

TCIII

[Thomas Coyle]

Hi Chris,

To be able to use Cooper's "IR_receiver_rosserial_ircode.ino" you will need to install the Rosserial Arduino Library in the Arduino IDE Library.

To do so, open the Arduino IDE Library Manager and type rosserial in the search window. It will bring up the Rosserial Arduino Library to install.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

After installing the Rosserial_Arduino_Library in the Arduino IDE you will discover the following:

If you try to compile my 'IR_Receiver_38k' code (after adding back in the ROS commands from Cooper's 'IR_receiver_rosserial_ircode.ino') with the Arduino IDE, you will find that it will not compile without error.

Go to this link and follow the instructions to fix ros/msg.h in the Rosserial_Arduino_Library.

Then recompile my updated 'IR_Receiver_38k' code and it will compile without errors.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

After compiling the  'IR_Receiver_38k' code with the additional ROS commands from Cooper's  'IR_receiver_rosserial_ircode.ino', the Arduino IDE indicated that the addition of the ROS commands really bumped up the amount of stored and dynamic memory in use.

So I compiled Cooper's 'IR_receiver_rosserial_ircode.ino' code and after the compilation, the Arduino IDE indicated that his program had just about maxed out the dynamic memory storage and there could be problems with execution of the that particular code.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Bump?

I have the docking station hardware all ready to go so all I need is your input on Cooper's 'mybot_autodocking' ROS code integration in the ArloBot software.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I have attached the IR Receiver Assembly to the front of the ArloBot Chassis.

I am now working on a carboard mockup of the docking station with the IR Transmitter mounted at the same level from the floor as the IR Receiver Assembly on the ArloBot.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I plan to have a Picam v2 mounted above the IR Transmitter on the docking station so I can view the progress of the ArloBot as it approaches and moves into the docking station.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Once we get the docking station/ArloBot function operational, maybe you can replace "unplug" in the webui with "dock/undock" for anyone who implements the docking station?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Will you be available this weekend?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Will you had a chance to review Wenny Cooper's "mybot_autodocking_irCode.cpp" ROS code this week?

I have the IR Receiver attached to the front of the ArloBot and the IR Transmitter mounted on a mockup of a docking station so I am already to test out the integration of Cooper's "mybot_autodocking_irCode.cpp" in the ArloBot software package.

Regards,

TCIII

[Christen Lofland]

Sorry about that. I gave it a brief look and realized that it was more work than I was expecting. The author didn't really create  full ROS "node". I will try to look at it this week and bundle it up into something that can be run with ROS.

[thomasco...@gmail.com]

Hi Chris,

That sounds great.

I was hoping that Cooper had done the heavy lifting for us, but something is better than nothing for a start.

I am ready to test whenever you have the code ready to test.

Do you think that we should open an "enhancement" issue on your github to track the testing progress?

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

Since you appear to have originally leveraged off of the TurtleBot software to create your ArloBot ROS software package, there are a number of TurtleBot articles concerning autodocking:

Link 1

Link 2     # Similar to Cooper's design, but uses three IR sensors. Link to project on YouTube video is broken. The robot chassis, though smaller than the Arlo, is interesting in its design.

Regards,

TCIII

[Christen Lofland]

The turtle bots were always based on vacuum cleaner chassis, so they had the benefit of usually working with an existing dock and hardware already set up to mate up and connect power.

I was very close to buying one of those turtle bots in 2016, but the Arlo platform lured me into rolling my own setup because it could hold so much weight. Looking back, I would have been fine with the turtle bot, but I learned so much.

Link 1, with the camera bases setup looks pretty cool, and he probably has a more complete code example. I do think dedicated IR sensors or docking would be cheaper than a dedicated camera, but maybe not?

The difficulty with so many of these ROS projects found online is that they are done as "research studies" in which they attempt to solve one very discreet problem. They don't consider how to integrate it well into a robot that does other things. They also are often submitting this for some college exam, so they only have to demonstrate a certain level of functionality, not actually have it work every day in random locations.

Anyway, regarding ROS and auto docking, this is how I see it working:

1. You have a map in ROS, and a "way point" set up for "near" the charger and "facing" it (your IR sensor seeing it, whether that is forward or backward). This part you already have, you can make a map, set way points, and tell it to go there. You can do it via the Web Interface or Rviz, and there are also command line scripts to do it in the setup.

2. The hardware that can "see" the dock, tell you if you need to rotate right or left and when to proceed. I think you have this now right?

3. The "glue" here is that code needs to publish "twist" messages on a ROS Topic. This is exactly what the keyboard, joystick, or Slam Toolbox do now. We just add one more publisher to the mix, and it gets its input from your hardware in step 2 and outputs twist messages.

I THINK that the code you referenced before does that, but they didn't wrap it up with ROS launch files, so I have to make those. There is also a "ROS Node" that basically "mixes" (prioritizes really) the various topics talking to the robot (How does it work to have  joystick, web site input, and Slam Toolbox all sending twist messages?! Well, there is a mixer!) that I need to add this topic to so that the messages get to the robot when they exist.

I just need to sit down and do it. Make the launch files, test it, add it to the mixer, and then publish the updates so you can test them.

This is also kind of a distraction from getting the Roboclaw controller working with ROS. :)

[thomasco...@gmail.com]

Hi Chris,

Yes, I have all of the IR System built and ready to go.

Since RoboClaw Motor Driver Modules will not be available until the middle of this year, you might want to complete the ROS docking code package so I can do testing until the RoboClaws become available?

Regards,

TCIII 

[thomasco...@gmail.com]

Hi Chris,

The video in Link 2 shows a very interesting approach to the docking station/robot electrical contact design and method of engagement that I might attempt to duplicate.

Regards,

TCIII

[Christen Lofland]

I started a Github issue to track this, as I'm getting lost in the group discussion.

Can you look at https://github.com/chrisl8/ArloBot/issues/198#issuecomment-1042350414

Follow the setup instructions, and the test instructions, and post there what you find.

Thank you.

[thomasco...@gmail.com]

Hi Chris,

I will post all future docking station comments and observations on your github.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I have posted a comment on github issue #11 as to some questions about the IR Receiver testing.

Regards,

TCIII

[thomasco...@gmail.com]

Hi Chris,

I have posted the results of testing both the 'mybot_autodocking.cpp' and 'mybot_autodocking_irCode.cpp' code on enhancement issue #198.

There still appears to be the need for the Autodocking Node to communicate with the Arduino IR Receiver USB output.

Regards,

TCIII 

[thomasco...@gmail.com]

Hi Chris,

This is why Wenny Cooper never completed his Autodocking Design: Andbot

Regards,

TCIII