Development of an Educational Omnidirectional Mobile Manipulator with Mecanum Wheels ^†

Abstract

The developed omnidirectional mobile manipulator is an educational omnidirectional mobile manipulator that utilizes the Raspberry Pi Pico W and is programmed in Python. It is designed to enhance STEM education by providing an interactive environment for studying robotics, sensor integration, and programming techniques. The robot is built on an off-the-shelf chassis equipped with Mecanum wheels and a robotic arm actuated by servo motors. As part of this project, the control electronics were designed and implemented to enable seamless operation. While the platform allows students to program the robot as part of the STEM curriculum, our base software solution, developed in Python, provides control of both the mobile base and the robotic arm via a web interface accessible through the robot’s Wi-Fi hotspot.

1. Introduction

The educational omnidirectional mobile manipulator (Figure 1) functions as a mobile robotic platform designed for educational applications in science, technology, engineering, and mathematics (STEM). This study presents a complete redesign of the robotic system’s electronics coupled with new, enhanced software solutions tailored for a designated programming platform [1]. The proposed alterations facilitate a diverse array of instructional methodologies targeting electronics, programming, and broader STEM concepts [2,3,4].

A fundamental aspect of STEM education is the practical demonstration of hardware integration, particularly through the use of educational mobile robotics and the associated robot control software [5]. This article examines the kinematic capabilities of the educational mobile manipulator, including robotic locomotion, arm articulation, object manipulation (gripping), transportation, and precise object placement within designated target areas. Furthermore, it details the software interfaces that manage the execution of the robot’s operational commands, along with the programmed sequences that synchronize the movements of the mobile unit and the functionality of the robotic arm.

2. Hardware

The robotic system, referred to as the educational omnidirectional mobile manipulator (Figure 2), employs the Raspberry Pi Pico W (Pololu Robotics and Electronics, Vegas, NV, USA) as its central processing unit [6,7]. This microcontroller facilitates communication with various peripheral components, thereby orchestrating the control of motors and sensors within the system. The architecture comprises several integral components, elaborated upon below:

• Raspberry Pi Pico W: The core microcontroller responsible for executing the control algorithms and managing data flow between the system components. It incorporates wireless networking capabilities, enabling remote control and monitoring.
• Motor Drivers: These components interface between the microcontroller and the motors, allowing the Raspberry Pi Pico W to control motor speed and direction via pulse width modulation (PWM) signals.
• DC/Servo Motors: Actuators that provide locomotion to the omnidirectional mobile manipulator. The specific type of motors employed is contingent on the required operational functions, such as movement and steering.
• Sensors: Various sensors are integrated into the robot to enhance its environmental awareness and capability. These may include the following:

  ○

  Ultrasonic Sensors: Utilized for distance measurement and obstacle avoidance.

  ○

  Infrared Sensors: Employed for line following or proximity detection.

  ○

  Camera Module (if applicable): Provides visual input for advanced capabilities such as image processing or object recognition.

• Power Supply: A dedicated power source that delivers the necessary voltage and current to the Raspberry Pi Pico W and the motor drivers, ensuring consistent operation throughout the system’s runtime.
• Communication Interfaces: These may include I2C, SPI, or UART communication protocols that facilitate data exchange between the Pico W and peripheral components, ensuring seamless coordination of the system’s functions.

In summary, the robot is a modular and versatile robotic platform that integrates various electronic components, allowing for diverse applications in robotics, automation, and educational environments.

Educational manipulator hardware architecture combines energy-efficient and functional modules that enable high maneuverability, autonomy, and interactivity [8,9,10]. The integration of Raspberry Pi Pico W, PCA9685, and Pico-Motor-Driver provide flexible control, and the ultrasonic sensor makes the robot more intelligent and adaptable in different environments. To provide stable and continuous power sRaspberry Pi Pico W, PCA9685upply, the mobile manipulator uses two Li-ion batteries (2600 mAh) to ensure autonomous operation. A DC-DC converter provides stable power supplied to some of the modules.

Table 1. Educational omnidirectional mobile manipulator hardware components.

№	Component	Picture	Features	Main Features
1	Raspberry Pi Pico W		Main microcontroller, which controls all peripherals and communication	ARM Cortex-M0+; 133 MHz; Wi-Fi module I2C; SPI and PWM interfaces; low power consumption
2	Pololu 5 V, 2.5 A Step-Down Regulator (D24V25F5)		A step-down regulator provides a stable 5 V supply for the components	Input voltage: 6–38 V; output voltage: 5 V; maximum current: 2.5 A; up to 90% efficiency
3	Pico-Motor-Driver (Waveshare)		Driver for 4 DC motors, used for drive control	Operating voltage: 6–12 V; 12-bit PWM control; H-bridge drivers TB6612FNG; I2C communication
4	PCA9685 (16-channel PWM controller)		Controls the servomotors of the manipulator via hardware PWM control	Up to 16 PWM channels; I2C interface; 12-bit precision; 5 V power supply
5	MG90S servomotors (3 pieces)		Control of the robotic arm—rotation, lifting, and gripping	Operating voltage: 4.8–6 V; torque: 2.2 kg/cm; metal gears
6	Ultrasonic CS100A		Distance measurement and obstacle detection	Range: 2–500 cm; power supply: 3.3–5 V; TRIG/ECHO interface
7	Li-Ion batteries (2 × 2600 mAh)		Provide autonomous power to the robot	Total capacity: 5200 mAh; high energy density; rechargeable

shows the hardware components used on the educational omnidirectional mobile manipulator.

The I²C bus configurations of the Pico W is shown in

Table 2. The Pico W I²C bus configurations.

Source	Pico W (GPIO pin)	Pico-Motor-Driver (pin)
SDA (serial data)	GP20	SDA
SCL (serial clock)	GP21	SCL
VCC (power supply)	5 V	VCC (pin 39)
GND (mass)	GND	GND (pin 38)

.

Role of Pins in Communication and Power Supply:

• SDA (Serial Data Line—GPIO 20 pin) and SCL (Serial Clock Line—GPIO 21 pin):

The SDA and SCL pins facilitate I²C (inter-integrated circuit) communication between the Raspberry Pi Pico W and the motor driver. The SDA line is responsible for transmitting data, while the SCL line provides the necessary timing for synchronizing data transfers. This synchronous communication protocol allows for efficient data exchange and control of the motor functions. Proper connectivity of the SDA and SCL lines is critical; misconfiguration will inhibit effective communication, resulting in the motor driver failing to respond to commands issued by the Raspberry Pi Pico W, thereby impairing system functionality.

• GND (Ground) and VCC (Voltage Common Collector):

The GND and VCC pins establish the required electrical connection between the Raspberry Pi Pico W and the motor driver. The GND pin serves as the reference point for the voltage levels within the system, ensuring a common ground for accurate signal transmission. The VCC pin provides the necessary voltage supply to the motor driver, enabling its operation. These connections are crucial for ensuring that both devices operate within their specified voltage ranges and that the signals are interpreted correctly. Inadequate or improper connections may lead to insufficient power supply or ground reference issues, potentially resulting in system malfunction or device damage.

3. Software Architecture of Educational Omnidirectional Mobile Manipulator

The educational omnidirectional mobile manipulator software architecture consists of several integral modules, all implemented in Python 3, which facilitate essential functionalities including inter-module communication, motor control, robotic arm manipulation, and user interaction through a web interface [8,9,11]. Each of these modules interacts cohesively to enable the seamless operation of the robotic system. The communication module ensures efficient data exchange between components, while the motor control module orchestrates the precise movements of the robotic platform. The robotic arm control module is responsible for the manipulation and positioning tasks, and the web interface serves as an accessible platform for user engagement and system monitoring. Together, these modules form a robust framework for the deployment and operation of the educational mobile manipulator platform. The main software modules developed for the educational manipulator are shown in

Table 3. Basic Python programming modules.

Module (File)	Feature
main.py	The startup script that initializes the robot and manages Wi-Fi
robot_main.py	Responsible for creating the Wi-Fi access point and handling HTTP requests
index.html	Web interface to control the movement of the robot base and robotic arm via browser
robot.py	Master class for robot control including motion methods
robot_motors.py	Class for controlling wheel motors via PCA9685 and H-bridge driver
robot_arm.py	Robotic arm control class using PCA9685 to control the servo motors
pca9685.py	Driver for PCA9685 to generate PWM signals for motors and servomotors

.

Module Structure and Communication Overview:

• Motor Control:
  The robot_motors.py file utilizes the PCA9685 to manage four DC motors.

  ○

  The motors are controlled via the I²C interface, specifically using GPIO20 (SDA) and GPIO21 (SCL).

  ○

  This module includes a motor driver class featuring methods to start, stop, and adjust the speed of the motors.

• Robotic Arm Control:
  The robot_arm.py file is responsible for controlling three MG90S servomotors through the PCA9685.

  ○

  Control is executed via I²C communication.

  ○

  The module supports smooth movement of the servomotors through the smooth_move_servo() method.

  ○

  It also includes a function to reset the arm (reset_servos()).

• Wi-Fi Communication and Web Interface:
  The robot_main.py file creates a Wi-Fi access point and manages requests from index.html.

  ○

  Upon startup, the mobile manipulator creates an SSID, i.e., “robot”, allowing users to connect.

  ○

  When users interact with the web interface by pressing a button, an HTTP request is sent to the robot.

  ○

  Movement and robotic arm control can be performed through sliders and buttons in the web interface.

The software workflow consists of the following stages:

• Power On and Initialization: The main.py script initializes robot.
• Wi-Fi Network Setup: The robot_main.py script creates the Wi-Fi access point (SSID: “robot”).
• Web Control Interface: The index.html file provides a graphical control panel for user interaction.
• Command Processing: The educational omnidirectional mobile manipulator processes requests from the web interface through robot_main.py.
• Motion and Manipulation: The robot.py script invokes the motion and control methods for the robotic arm.
• Connection Termination: The Wi-Fi connection can be terminated when the user logs off.

The Python modules (files) associated with the omnidirectional mobile manipulator workflow are shown in

Table 4. Python modules related to the workflow.

№	Action	File
1.	Initialization of the system	main.py
2.	Setting up a Wi-Fi access point	robot_main.py
3.	Interactive web page interface	index.html
4.	Accept and process commands	robot_main.py
5.	DC motor control	robot.py
6.	Arm control	robot_arm.py
7.	Execute commands from the web interface	robot_main.py
8.	Execute commands from the web interface	robot.py

.

The presented architecture facilitates a dynamic and efficient management framework for the educational omnidirectional mobile manipulator, enabling the augmentation of its functionality through the integration of novel modules and enhancements.

Figure 3 illustrates a block diagram representing the primary components of the program. The educational manipulator management program operates on a fundamental iterative loop, which methodically processes incoming requests from clients interfacing through the robot’s Wi-Fi access point. Each request may encapsulate instructions for platform control, motor operation, or manipulation of the robotic arm, thereby enabling versatile interaction with the robotic system. This design paradigm ensures the seamless execution of commands and adaptive response to user inputs, thereby enhancing the overall operational efficacy of the educational manipulator.

The primary software snippets illustrating the educational omnidirectional mobile manipulator workflow are as follows:

Power-Up and Initialization.

File: main.py.

Upon startup, the robot loads the essential modules and initializes its control systems.

import sys.

sys.path.append(‘/Robot’).

from robot_main import * #—the main control module.

We create a Wi-Fi network. File: robot_main.py. Upon startup, the omnidirectional mobile manipulator creates a Wi-Fi access point with the SSID “robot”, allowing users to connect seamlessly (Figure 4). Robot control—web interface. File: index.html. HTML interface with motion buttons (Figure 5). HTTP request processing and command execution. File: robot_main.py. When a button is pressed, the robot receives an HTTP request and executes the corresponding movement (Figure 6).

Motion Control. File: robot.py. Motion functions by controlling the four DC motors (Figure 7). Control of the robotic arm. File: robot_arm.py. Control of the three servomotors via PCA9685. (Figure 8) Handling hand control requests from the web interface. File: robot_main.py. When the values of the web interface sliders are changed, a request is sent to the robot (Figure 9).

Session completion and shutdown.

File: robot.py.

A method to shut down all engines when a job is completed.

def stop_all_motors(self):

self.m.StopAllMotors().

4. Experimental Studies

To assess the functionality of the robot (see Figure 1), a series of experiments were conducted to evaluate its control precision, movement capabilities, and object manipulation proficiency. The robot was connected via the Wi-Fi access point (AP) provided by software running on the Raspberry Pi Pico W.

The robot’s configuration for these experiments includes the following:

• A chassis equipped with four Mecanum wheels for enhanced multi-directional mobility.
• A manipulator featuring three axes of motion and a gripper.
• A Raspberry Pi Pico W serving as the central controller.
• A battery power supply alongside drivers for DC and servo motors.
• A web interface for controlling the robot (illustrated in Figure 2).

The experimental setup involved a smooth, flat, and rigid surface that provided stability for the robot’s movements. This environment included a platform for placing the objects the robot was to grasp, as well as a marked circle indicating the designated target location for placement.

Two experiments were conducted in this study:

• The first experiment assessed the robot’s motion control to verify the system’s operational accuracy.
• The second experiment focused on the robot’s capacity to grasp, transport, and accurately place objects within the designated target area.

Experiment 1.

Control and Movement of the Robot.

Figure 10 illustrates the various movements of the manipulator, which can be managed through the robot’s web interface. This interface features sliders dedicated to controlling the three primary axes of the manipulator:

• Base: Rotates the manipulator’s base to the left or right.
• Arm: Adjusts the arm’s position vertically, allowing for upward or downward movement.
• Gripper: Regulates the opening and closing mechanism of the gripper.

Upon initialization, the default position for all three axes is set to 90 degrees.

Table 5. Control of gripper positions.

№	Image	Handler Movement	Control Used (Slider)	Control Direction
1	Row 1, Left	Start position of the gripper	-	Set to center
2	Row 1, Middle	Rotate to the right and slightly lower	Base, Arm	Slid right, slid left
3	Row 1, Right	Turn left and lift slightly	Base, Arm	Slid left, slid right
4	Row 2, Left	Closing the gripper	Gripper	Slid right
5	Row 2, Middle	Opening the gripper	Gripper	Slid left
6	Row 2, Right	Rotation towards center	Base	Slid to center

outlines the sequence of experiments conducted to control the gripper, while Figure 11 displays the corresponding positions of the gripper throughout these experiments.

To manage the platform’s movements, the experiment utilized a system of random button selection. Examples of these commands, accompanied by actual button images, can be seen in Figure 11.

After each movement, the robot can be redirected without the need to press the “STOP” button.

Experiment 2.

Capture, transport, and placement of an object at a target location.

This experiment is designed to assess the mobile robot’s capability to grasp, transport, and accurately place small objects at designated target locations. Figure 12 provides visual documentation of the experiment. The robot is controlled via a web interface, allowing for remote operation of both its movement and the manipulator.

Table 6. The experimental stages for object manipulation.

№	Stage	Action of the Robot	Comment
1	Approach to the site	The robot approaches the platform where the yellow 3D-printed object is located	Control is undertaken manually via the web interface
2	Site capture	The manipulator positions itself above the object and grips the 3D-printed element	The servo mechanism gently closes the grip so as not to damage the object
3	Transport of the object	The robot moves towards the target area marked with a black circle	Movement is performed using Mecanum wheels for flexible maneuvering
4	Placement of the object	The manipulator positions the object over the target marker (black circle) and releases it	The positioning accuracy is controlled to be placed correctly
5	Completion of the experiment	The robot stops after releasing the object	The user visually confirms successful execution

shows the experimental stages.

5. Conclusions

The experiments conducted validated the core functionality of the robot and its proficiency in executing the designated tasks. Testing revealed several key advantages, as well as some limitations that could be addressed in future iterations.

Table 7. Positive results of the experiments.

№	Advantage	Description	Benefit
1	Successfully implemented hardware and software	The robot is fully functional and works as expected	Suitable for testing, training, and further development
2	Connecting to the robot	The robot software creates a Wi-Fi access point, allowing for easy connection from a computer or mobile device	Quick and convenient robot startup without additional infrastructure
3	Control	Control is accessible via a web interface, with no complicated configuration	Easy to use even by novice users
4	Control flexibility	The robot can be controlled remotely via a PC without the need for additional controllers	Facilitates experimentation and learning by allowing quick changes in movements
5	Suitable for teaching students	Free source code and documentation make it easy to adapt the project for educational purposes	Suitable for STEM learning and developing algorithmic thinking
6	Functionality as expected	Basic object movement and manipulation work robustly	The reliability of the mechanics and software has been confirmed
7	Good maneuverability	Mecanum wheels provide freedom of movement in all directions	Suitable for complex manipulation tasks and navigation in confined spaces

and

Table 8. Identified deficiencies from the experiments.

№	Disadvantage	Description	Potential Risk
1	Lack of real-time management	Commands are not executed instantaneously due to communication delays	Incorrect or delayed robot responses
2	Risk of communication breakdown	Control relies on Wi-Fi, which can be affected by interference	The robot may remain in the last submitted state without control
3	Delays in user responses	Control via sliders and buttons requires manual actions by the operator	Reduced speed in emergency situations or fine tuning
4	Limitations of gripper mechanics	Excessive tightening can cause servomotor failure	Lack of overstretch protection, possible blockage
5	Lack of autonomous adjustments on the fly	The robot does not self-correct position errors	Lower accuracy when grasping and placing objects
6	Risk of hardware wear	Lack of smooth transitions in motor control	Faster mechanical wear and damage in the long term

provide a summary of the positive outcomes and identified shortcomings of the robot.