sra-vjti / wall-e Goto Github PK

Branch: saharsh/logger

logI(TAG_SRA, "%s", "Configuring pwm_config...\n");

@saharshleo, the '\n' character is redundant here, as the logger function handles it itself.

• All main.c needs to be ported to esp-idf v5.1
• Testing of each code with latest SRA board needs to be performed

Ref- https://github.com/SRA-VJTI/Wall-E/actions/runs/3923376894/jobs/6706946550

• Fix linting

../main/self_balancing.c: In function 'balance_task':
../main/self_balancing.c:92:23: error: lvalue required as left operand of assignment
    PITCH_ANGLE_OFFSET = setpoint;
                       ^
../main/self_balancing.c:107:9: error: 'pitch_error' undeclared (first use in this function); did you mean 'pitch_cmd'?
     if (pitch_error > 1)
         ^~~~~~~~~~~
         pitch_cmd
../main/self_balancing.c:107:9: note: each undeclared identifier is reported only once for each function it appears in

git blame shows, @OSSome01 submitted a PR for this, please fix the issue in self_balancing.c

This task should be done by the second years who will be taking up the task to teach concepts, rather than third years.

1. weighted sum should be '+=' not '='
   https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/d1ef385a9fbfd6569072ba660fd9bd2ea937169e/line_following/main/main.c#L76-L83

2. and if condition should be line_sensor_readings.adc_reading[i] < 400

3. Your map function is giving zero value, fix that.
   https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/d1ef385a9fbfd6569072ba660fd9bd2ea937169e/line_following/main/main.c#L52-L54

4. Not issue but if possible create macro for 400 (black margin) and 2000(white margin)
   https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/d1ef385a9fbfd6569072ba660fd9bd2ea937169e/line_following/main/main.c#L119-L120

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/e941ee6b972d7a36fa6d28c94547fff3bf159392/components/sra18/sra18.c#L66

Are these switches or buttons, buttons usually need some debouncing..

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/599e674feaee17bf6a6cf1048582e299eddff551/8_self_balancing/main/self_balancing.c#L53

in line number 53 pitch_error_cumulative is multiplied with dt to get pitch_area but it should have been pitch_area+=pitch_error*dt. So why are we using multiplying pitch_error_cumulative with dt?

TLDR;
We need to do summation (error*dt), but the current code does summation(error)*dt.

Submitted a possible fix in (#64)

@luke2103 add this at the end of while loop

vTaskDelay(100 / portTICK_PERIOD_MS);	

and this will be "MOTOR_A_0 Backward", right ?

		ESP_LOGI(TAG_SWITCH_CONTROL_MOTOR, "MOTOR_A_0 FORWARD");	
	}
	else if(read_switch(SWITCH_2))
	{
		set_motor_speed(MOTOR_A_0, MOTOR_BACKWARD, 80);
		ESP_LOGI(TAG_SWITCH_CONTROL_MOTOR, "MOTOR_A_1 BACKWARD");

Assign numbers and label them according to how one would process in workshop
So

1. led_test
2. Some other code
3. Some other code but not the above on

If by any chance you face an error like cbor library not found once the installation completes, try switching to
release/v4.2, origin/release/v4.2 branch by

• cd %IDF_PATH%
• git checkout origin/release/v4.2

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/8404030925bf08c08f5ffae1843dd1207e65676a/pwm_test/main/pwm_test.c#L9-L10
instead of motor driver header, include sra board header.

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/8404030925bf08c08f5ffae1843dd1207e65676a/pwm_test/main/pwm_test.c#L49-L50
@OSSome01 you have defined macro named MODE so replaced 'mode' by MODE.

In the dev branch , under 6_PWM , the file main/pwm_test.c
GitHub Link
Why are the following headers included ?

//C Headers
#include <stdio.h>
#include <math.h>
#include <time.h>
#include <string.h>
#include <stdlib.h>

I have tried commenting the headers and building .

Following is my directory git status which shows what I commented -

satacker@ubuntu:~/Desktop/Wall-E_v2.2-beta/6_PWM$ git diff HEAD
diff --git a/6_PWM/main/pwm_test.c b/6_PWM/main/pwm_test.c
index 38a9db0..3575025 100644
--- a/6_PWM/main/pwm_test.c
+++ b/6_PWM/main/pwm_test.c
@@ -1,9 +1,9 @@
 //C Headers
-#include <stdio.h>
-#include <math.h>
-#include <time.h>
-#include <string.h>
-#include <stdlib.h>
+// #include <stdio.h>
+// #include <math.h>
+// #include <time.h>
+// #include <string.h>
+// #include <stdlib.h>
 
 
 //Components
diff --git a/components/sra-board-component b/components/sra-board-component
index 9f28700..3be72ec 160000
--- a/components/sra-board-component
+++ b/components/sra-board-component
@@ -1 +1 @@
-Subproject commit 9f28700759ac816660c18859d65303d6540e8732
+Subproject commit 3be72ec615845e355a1b7eb8b8ef5d4029ae217b

Build Output
Skipping directly to the end -

Scanning dependencies of target pwm_test.elf
[ 99%] Building C object CMakeFiles/pwm_test.elf.dir/project_elf_src_esp32.c.obj
[100%] Linking CXX executable pwm_test.elf
[100%] Built target pwm_test.elf
[100%] Generating binary image from built executable
esptool.py v3.1-dev
Generated /home/satacker/Desktop/Wall-E_v2.2-beta/6_PWM/build/pwm_test.bin
[100%] Built target gen_project_binary
[100%] Built target app
Executing action: all (aliases: build)
Running make in directory /home/satacker/Desktop/Wall-E_v2.2-beta/6_PWM/build
Executing "make -j 10 all"...

Project build complete. To flash, run this command:
/home/satacker/.espressif/python_env/idf4.3_py3.8_env/bin/python ../../../esp/esp-idf/components/esptool_py/esptool/esptool.py -p (PORT) -b 460800 --before default_reset --after hard_reset --chip esp32  write_flash --flash_mode dio --flash_size detect --flash_freq 40m 0x1000 build/bootloader/bootloader.bin 0x8000 build/partition_table/partition-table.bin 0x10000 build/pwm_test.bin
or run 'idf.py -p (PORT) flash'

adc_chars = calloc(1, sizeof(esp_adc_cal_characteristics_t));

https://github.com/SRA-VJTI/sra-board-component/blob/a11b18b2f91b80a4b7a0e5ae5a1d5f1f89470699/src/adc.c#L50
I guess this calloc should be freed.

Code was referred from ESP32 ADC example:
https://github.com/espressif/esp-idf/blob/master/examples/peripherals/adc/main/adc1_example_main.c
They haven't done anything for calloc, @VedantParanjape @laukik-hase kindly check once.

MPU read function defined as:
esp_err_t read_mpu6050(float *euler_angle, float *mpu_offset);
(https://github.com/SRA-VJTI/sra-board-component/blob/2b96f498ccd8b1a19d42bc7d09e1c36d513d6b4e/include/mpu6050.h#L130-L137)

But in the MPU code:
https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/0f3ffc9b890600030094a3bc40c7a026c6a9fcd8/mpu_test/main/main.c#L24

fewer arguments and @laukik-hase, @OSSome01 check the MPU part in self-balancing code on the dev branch.
In that code also MPU is not initializing.

The line following webpage doesn't work well on firefox.

Currently the latest ESP-IDF release version is 4.4.1. All the firmware of Wall-E should be updated to be able to compile with that IDF.

Windows Testing Task

Setup

• ESP-IDF v4.2
• Installed using esp-idf-tools-setup-2.4.exe
• idf-python 3.8.7

Testing

• Wall-E_v2.2-beta\6_PWM at origin/dev

Structure

C:\Users\shrey\Desktop\Wall-E_v2.2-beta>tree .
Folder PATH listing
Volume serial number is 005F0045 02C1:3324
C:\USERS\SHREY\DESKTOP\WALL-E_V2.2-BETA
├───1_led_blink
│   └───main
├───2_LSA
│   └───main
├───3_MPU
│   └───main
├───4_switch_controlled_motor_normal
│   └───main
├───5_switch_controlled_motor_parallel
│   └───main
├───6_PWM
│   └───main
├───7_line_following
│   ├───frontend
│   └───main
│       └───include
├───8_self_balancing
│   ├───frontend
│   └───main
│       └───include
├───components
│   └───sra-board-component
│       ├───.github
│       │   └───workflows
│       ├───assets
│       ├───docs
│       │   ├───def
│       │   ├───search
│       │   └───xml
│       ├───examples
│       │   ├───bar_graph
│       │   │   └───main
│       │   ├───lsa
│       │   │   └───main
│       │   ├───motor_driver_normal
│       │   │   └───main
│       │   ├───motor_driver_parallel
│       │   │   └───main
│       │   ├───mpu6050
│       │   │   └───main
│       │   ├───servos
│       │   │   └───main
│       │   └───switches
│       │       └───main
│       ├───include
│       └───src
└───windows_instructions

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/06084b637ec267d9d5588e426e3d7b3672a4b062/8_self_balancing/main/self_balancing.c#L48

Update the firmware following 7_line_following wrt new design of LSA, also taking into account that the new LSA outputs the raw ADC data in a reverse manner when compared to the current LSA in use.

in WallE v2.1 we had an example code of bot following line while self balancing as well. So it would be good if for further wallE workshops we can add it here as well.

• Change the Wall-E 2.2 to Wall-E 2.3
• Include Oled images
• Change board image

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/e941ee6b972d7a36fa6d28c94547fff3bf159392/self_balancing_with_line_following/main/self_balancing_with_line_following.c#L32

float kd = 9;
float output = kd * 0.5;

output could be 0..

Prefer to define all floats in this manner float kd = 9.0f

https://github.com/SRA-VJTI/sra-board-component/blob/main/src/bar_graph.c#L141

https://github.com/SRA-VJTI/Wall-E_v2.2-beta/blob/e941ee6b972d7a36fa6d28c94547fff3bf159392/self_balancing_with_line_following/main/self_balancing_with_line_following.c#L147
I see from the youtube that line following is quite slow.

I think it can be made to go faster by bounding pitch command to a different limit on self balancing while line tracking:

-10 deg (pitch forward) to -3 deg (still pitch forward) could result in better line tracking...
Usually, you'd only want to pitch lesser (-3 deg) when you are on a turn or if the heading of the bot is not aligned with the line. Bounding from -10 deg to 0 degree makes the bot stop at one point. Any opinions?

Change function name from init_line_sensor() to enable_line_sensor()

https://github.com/SRA-VJTI/sra-board-component/blob/30ce480aa9a16ecdb565f9ffa874e0b78720a1d9/src/lsa.c#L5
https://github.com/SRA-VJTI/sra-board-component/blob/30ce480aa9a16ecdb565f9ffa874e0b78720a1d9/include/lsa.h#L31

• Adding a separate folder for the Line and Self balancing
• Should be written according to the latest ESP-IDF (v5.1)
• Testing the code with the latest hardware

Deploy a simple action that verifies install script across Ubuntu-(whichever we are going to use), Windows and mac osx.

GitHub has such runners available.

CC: @SAtacker @Jamm02 @gautam-dev-maker

As this repo is growing and will be used by a lot of different people, commenting code obviously makes it easy for new collaborators.
But even as collaborators, it should not always be necessary for us, to navigate into the existing code just to look at the comments. This could be avoided if we have some sort of documentation for our code. We do realize that none of us will really like to put in dedicated effort just to build documentation.
The easiest workaround for this is to use tools that can build documentation straight from the comments we write in the code. (I am calling it a workaround, but it is also the general practice)
Examples of projects that use these tools

Some of the tools that could be tried:
Doxygen
Sphinx
DOC++
I looked at some of the code and you guys have done good work with maintaining comments. Just formatting the existing comments should give us enough to evaluate whether this is a practice we could take up. It should be a fairly straightforward task. And it will probably take 3 or 4 hours once one of us actually gets down to formatting the source code.

While LSA Readme contains description of functions used, MPU Readme doesn't contain any mention of functions used.

such non-uniformity in README.md should be resolved.