pymycobot::get_angles may fail depending on current angle condition about pymycobot

[Eheech]

Would you mind uploading your codes ? We tried send_angles() and get_angles() multiple times , didn't have this issue , so if you don't mind, we'd like to test it with your code , thanks ~

[SidKoe]

I am not used to using GitHub so I attach changed code here.
Code I have changed is as below. It just DataProcessor::_process_received() only.
Changed lines is marked with command, "# CHANGED"
I also attach binary dump of transmitting data and comment following code. I hope it helps you to understand issue.

---- CODE:start ----
def _process_received(self, data, genre):
if not data:
return []

    data = bytearray(data)
    data_len = len(data)
    # Get valid header: 0xfe0xfe
    for idx in range(data_len - 1):
        if self._is_frame_header(data, idx):
            data_len = data[idx + 2] - 2
            if data_len > 0:
                cmd_id = data[idx + 3]   # CHANGED
                if cmd_id == genre:       # CHANGED
                    break                         # CHANGED
    else:
        return []

    data_len = data[idx + 2] - 2

    # compare send header and received header
    cmd_id = data[idx + 3]
    if cmd_id != genre:
        return []
    data_pos = idx + 4
    valid_data = data[data_pos : data_pos + data_len]

    # process valid data
    res = []
    if data_len == 12:
        for idx in range(0, len(valid_data), 2):
            one = valid_data[idx : idx + 2]
            res.append(self._decode_int16(one))
    elif data_len == 2:
        res.append(self._decode_int16(valid_data))
    else:
        res.append(self._decode_int8(valid_data))
    return res

---- CODE:end ----

Each line is divided by chunk of command or response.
Data start with servo related communication data and last chunk is response from M5atom, I thought.
Last servo response terms with check sum and it value is 0xfe.
Succeeding M5atom response start with 0xfe, 0xfe, header part.
Current code will take termination check sum code and first one byte of data as starting response of M5atom.
And pymycobot library will check command code at wrong data and fail parsing.
---- DATA:start ----
ff ff 01 04 02 38 02 be
ff ff 01 04 00 87 06 6d
ff ff 02 04 02 38 02 bd
ff ff 02 04 00 f1 03 05
ff ff 03 04 02 38 02 bc
ff ff 03 04 00 43 07 ad
ff ff 04 04 02 38 02 bb
ff ff 04 04 00 70 03 84
ff ff 05 04 02 38 02 ba
ff ff 05 04 00 11 08 dd
ff ff 06 04 02 38 02 b9
ff ff 06 04 00 f0 07 fe
fe fe 0e 20 0c f1 23 ab f9 8c 28 19 ff 6B 00 8c fa
---- DATA:end ----

[SidKoe]

P.S.
Well, I think you may repro more easier when you change angle of J6. I think last response is for J6 servo. Others may not cause this issue. If you have any question, please let me know.

[Eheech]

In pymycobot , it will filler out communication messages between Basic and Atom , only read Get_angles()

[Eheech]

When you get empty array, have you check Atom's status ? Is it 'yes' or 'no' ? and have you chosen ' minirobot -> transponder ' ?

[Eheech]

Maybe you can upload your code which you used to control mycobot , we'd like to test it with your code , thanks ~

[SidKoe]

I have not yet checked Atom's status. I am not sure which API is available to see Atom's status. But even if this issue happened mycobot can be controlled from pymycobot API; changing angle of joints, focus on or off servos. I believe this issue is not problem of Atom or Base. It will be handling problem of received data in pymycobot or something.
I am using transponder and set various angles of all joints. Most cases it works fine; get_angles() returns list of angles of joints.
BTW, it is difficult to provide code controlling mycobot as it is, because it includes many functions for another project. I will try to arrange something to provide for this issue.

[Eheech]

Are you using mycobot_280_M5 ? In transponder, press button A , it will return Atom's status in Basic.
You can use ' is_controller_connected() ' to check Atom's status , return data 1: connected 0: not connected -1: error .
And please update Atom and Basic's firmwares before you use mycobot .

[tongtybj]

Hi, I face the same with MyCobot Pro 320 M5.
My code is based on mycobot_ros.
The debug branch is https://github.com/tongtybj/mycobot_ros/tree/interface_bug_mycobot320pro_m5.

Code

You can try following command to check the bug:
roslaunch mycobot_communication mycobot_interface.launch

Then you can find following log result:

get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []
get angles: [-68.64, 129.37, -138.6, -36.65, 91.49, -0.35]
get angles: []
get angles: []

The related code is here:
https://github.com/tongtybj/mycobot_ros/blob/interface_bug_mycobot320pro_m5/mycobot_communication/scripts/mycobot_interface.py#L63-L64

Other tests:

• I also tried with MyCobot280 M5, but this bug never happened (the list never be empty).
• If I comment out
  https://github.com/tongtybj/mycobot_ros/blob/interface_bug_mycobot320pro_m5/mycobot_communication/scripts/mycobot_interface.py#L77-L78, this bug is also disappeared for MyCobot Pro 320 M5.

Firmware version:

• Mycobot Pro 320 M5: basic is v1.0; atom is v4.1
• Mycobot 280 M5: basic is v1.0; atom is unknown

[v-ghobadi]

get_angles()returns an empty array on the client while working with myCobot 280pi.