Abstract

Robotic automation is increasingly critical in modern industry; however, its adoption in small U.S. manufacturing facilities remains limited due to high costs and concerns regarding structural reliability. This paper presents the design, fabrication, and experimental validation of a lightweight, low-cost six-degree-of-freedom (6-DOF) industrial robotic arm intended for precise, low-payload applications such as light material handling and conveyor-based bottle sorting. Carbon fiber–reinforced polymer (CFRP) was selected as the primary structural material to achieve a high stiffness-to-weight ratio while maintaining a compact and efficient design. The robotic arm is engineered for seamless integration with a mobile platform, enhancing deployment flexibility and adaptability across diverse operational environments. Real-time control is achieved through an Android-based graphical interface that computes inverse kinematics and transmits joint commands to an onboard microcontroller via serial communication. The system enables precise control of joint position, velocity, and acceleration. Experimental results demonstrate accurate and reliable performance for small-scale industrial tasks. The mechanical structure was modeled and optimized using Fusion 360, while a user-friendly human–machine interface (HMI) developed in Android Studio enhances operational efficiency and workflow usability. Overall, the robotic arm offers a cost-effective and flexible solution for small manufacturing facilities, effectively bridging the gap between high-performance industrial robotics and lightweight, deployable automation platforms. Experimental results demonstrate reliable system performance, achieving an average trajectory completion time of 13.2 s with a variation of ±0.3 s across repeated trials. The robotic arm supports lightweight handling tasks with consistent repeatability and stable motion execution. Compared to conventional industrial robotic systems, the proposed design significantly reduces system complexity and cost while maintaining acceptable positioning performance for small-scale manufacturing applications.

1. Introduction

Industrial robotic manipulators have become integral components of modern manufacturing systems, significantly enhancing productivity, precision, and workplace safety by automating repetitive, hazardous, and labor-intensive tasks ^{1, 2}. Since their initial deployment in the 1960s, robotic arm technology has evolved substantially through advances in kinematics, control theory, sensing technologies, materials engineering, and embedded computing platforms ^{3, 4, 5}. Among the various manipulator configurations, six-degree-of-freedom (6-DOF) robotic arms remain the most widely adopted due to their high dexterity, extended reachability, and ability to perform complex positioning and orientation tasks within constrained workspaces ^{6, 7}.

Despite these technological advancements, the adoption of industrial robots in small-scale manufacturing facilities in the United States remains limited. High acquisition and maintenance costs, restricted scalability, limited customization options, and concerns regarding long-term structural reliability constitute major barriers ^{8, 9}. Consequently, many small manufacturers continue to rely on manual labor for lightweight and repetitive material handling operations, underscoring the growing demand for compact, low-cost robotic systems optimized for low-payload applications ^{10, 11, 12}.

Recent research has highlighted the benefits of lightweight robotic manipulators fabricated using advanced composite materials. Carbon fiber–reinforced polymer (CFRP) exhibits exceptional stiffness-to-weight and strength-to-weight ratios, enabling reduced inertial loads, improved dynamic performance, and enhanced energy efficiency ^{13, 14}. CFRP-based robotic structures achieve substantial mass reduction while preserving mechanical rigidity, making them particularly suitable for mobile platforms and space-constrained industrial environments ^{15, 16, 17}.

In parallel, mobile-device-based control architectures have emerged as viable alternatives to traditional industrial hardware interfaces. Modern Android platforms offer significant computational capability, low-latency communication, and intuitive graphical user interfaces, supporting real-time inverse kinematics computation, trajectory planning, and system monitoring ^{18, 19}. When integrated with microcontroller-based servo actuation, such architectures reduce system complexity, lower overall cost, and simplify programming and deployment, thereby improving accessibility for small and medium-sized manufacturing enterprises (SMEs) ^{20, 21, 22}.

However, the integration of lightweight composite manipulators, mobile-device-based motion control, and microcontroller-driven actuation into a unified industrial robotic solution remains relatively unexplored. Existing commercial systems from major manufacturers such as ABB, FANUC, and KUKA provide high precision and robustness but are often oversized, costly, and poorly suited for lightweight handling tasks in small-scale operations ^{23, 24, 25}.

This paper addresses these limitations by presenting the design and implementation of a customized 6-DOF robotic arm constructed primarily from CFRP and controlled via an Android-integrated interface. The system is optimized for low-payload industrial applications, including conveyor-based object sorting and light material handling, and is deployable in confined or mobile industrial environments. The architecture integrates computer-aided design (CAD), real-time inverse kinematics computation, and a user-friendly human–machine interface (HMI). The Android platform performs inverse kinematics calculations and transmits joint commands to an onboard microcontroller through a serial communication interface, enabling real-time control of joint position, velocity, and acceleration.

By combining lightweight composite materials, mobile-integrated control, and microcontroller-based actuation, the proposed robotic system provides a compact, cost-effective, and high-performance alternative to conventional industrial manipulators. This work bridges the gap between high-end robotic platforms and the practical automation needs of small-scale manufacturing facilities, contributing to broader adoption of robotic automation for lightweight industrial handling applications. Unlike conventional industrial robotic systems that rely on high-cost hardware and complex control architectures, the proposed system integrates lightweight CFRP-based structural design with an Android-driven control interface and low-cost microcontroller-based actuation. This combination enables a compact, flexible, and cost-effective robotic solution specifically tailored for small-scale manufacturing environments, where traditional automation systems remain economically impractical.

2. Methodology

The Denavit–Hartenberg (D–H) convention is used to relate the joint variables to the end-effector pose, with the corresponding parameters ²² listed in Table 1.

• θ_i: Joint angle (angle between X_i-1 and X_i along Z_i-1)

• d_i: Link offset (distance between X_i-1 and X_i along Z_i-1)

• a_i: Link length (distance between Z_i-1 and Z_i along x_i)

• α_i: Link twist (angle between Z_i-1 about X_i)

link's transformation matrix is computed as:

(1)

Given a desired end-effector pose , inverse kinematics (IK) computes the corresponding joint angles. For a spherical wrist configuration, the wrist center position is calculated as:

(2)

where represents the unit vector along the end-effector z-axis. Using geometric relationships, the first three joint angles are solved to position the wrist center. The remaining joint angles are then obtained from the wrist orientation. Due to the nonlinear nature of the IK equations, multiple valid solutions may exist, requiring appropriate selection based on joint limits and configuration constraints.

The workspace of the 6-DOF robotic arm is defined as the three-dimensional volume of all attainable end-effector positions subject to the mechanical limits of each joint. This workspace is influenced by joint reachability, dexterity provided by the last three wrist joints, and the presence of kinematic singularities. The workspace is visualized by uniformly sampling all joint variables within their allowable ranges and plotting the corresponding reachable end-effector positions, as illustrated in Figure 1.

The development of an Android-integrated control system for a six-degree-of-freedom (6-DOF) robotic arm using an Arduino platform was carried out through a modular and systematic methodology. The overall system architecture was divided into two primary subsystems: front-end and back-end, as illustrated in Figure 2. This separation enabled parallel development, improved system scalability, and simplified integration and testing.

The front-end subsystem focuses on user interaction and command generation, while the back-end subsystem is responsible for hardware control, signal processing, and execution of motion commands. Communication between the two subsystems is established through a wireless interface, enabling real-time control and monitoring of the robotic arm. This layered architecture reflects industry-standard practices in human–machine interface (HMI) and embedded robotic system design.

The front-end subsystem consists of an Android-based mobile application that provides an intuitive user interface for controlling the six-degree-of-freedom (6-DOF) robotic arm. The application enables users to issue motion commands, select predefined movement modes, and manually control individual joints. User inputs are translated into structured control commands and transmitted wirelessly to the back-end subsystem via Bluetooth communication.

The Android application was developed using Android Studio and referred to as “RASSLBot”, serves as the primary control platform for the system and is fully integrated with an Arduino UNO microcontroller, enabling reliable, low-latency communication for real-time robotic control. As shown in Figure 3, Figure 4, and Figure 5, the application offers two control modes: polar and Cartesian. Polar mode provides direct joint-level control, while Cartesian mode enables end-effector positioning in task space. In addition, the application supports a programming mode that allows users to record, save, and store industrial process motion sequences, which can be executed at any time in either manual or automatic operation modes. This multi-mode control architecture enhances operational flexibility and supports precise, repeatable robotic manipulation while remaining accessible to users with minimal technical background.

This section describes the mechanical, electronic, and software structures of the robotic system. The mechanical subsystem was designed using Autodesk Fusion 360 to ensure structural integrity, precision, and repeatability while supporting a variety of industrial manipulation tasks. The design considers link geometry, joint configuration, and load distribution to achieve reliable and stable motion performance.

The electronic subsystem integrates actuators, sensors, and control electronics to execute motion commands accurately. Servo actuators provide joint-level actuation, while the microcontroller processes control signals and sensor feedback to coordinate the overall operation of the robotic arm. The Android-based user interface functions as a communication bridge, translating high-level user inputs into low-level executable commands.

The end-effector is designed to emulate human-arm dexterity, with an emphasis on precise positioning and reliable load handling, enabling the system to meet the functional requirements of lightweight industrial applications

The robotic arm features six revolute joints, each providing one degree of rotational freedom, enabling full spatial manipulation of both position and orientation. As shown in Figure 6. The arm is divided into three main functional sections:

• Base and Waist (Joint 1): Enables rotation around the vertical axis (Z-axis), providing horizontal workspace coverage.

• Shoulder and Elbow (Joints 2 and 3): Allow vertical arm movement in the Y–Z plane, controlling end-effector height and reach.

• Wrist (Joints 4, 5, and 6): Provide orientation control along roll, pitch, and yaw axes for precise tool alignment.

Each joint is connected by a rigid link designed to optimize torque transmission, reduce weight, and maximize reach. Table 2 summarizes the design characteristics of each joint and link:

The links and frame of the robotic arm are manufactured from carbon fiber–reinforced polymer (CFRP) to achieve a high strength-to-weight ratio, reducing overall mass while maintaining structural rigidity. This material improves dynamic performance and maneuverability. High-precision rolling bearings at each joint provide smooth rotation, reduce friction, and minimize backlash, enhancing positioning accuracy.

The end-effector includes a modular mounting interface, allowing quick replacement of tools such as mechanical grippers, vacuum suction cups, or vision cameras, depending on the task. The arm is mounted on a rigid base that ensures mechanical stability and attenuates vibrations during motion. A complete 3D CAD model of the arm was developed, including exploded views and detailed assembly drawings to support manufacturing, assembly, and maintenance.

The electronic subsystem ensures precise motion control, reliable communication, and safe operation as shown in Figure 7. High-torque servo motors with integrated encoders are installed at each joint for closed-loop control. Encoders provide real-time feedback and support PWM signals, along with features such as overcurrent protection, thermal shutdown, and soft start for improved reliability and performance.

Arduino UNO executes inverse kinematics algorithms, coordinates joint movements, and processes sensor data for accurate end-effector positioning. A regulated power supply delivers stable voltage and current to all components, while Bluetooth modules (HC-06) enable wireless control and monitoring via the Android user interface. This integrated architecture provides responsive, high-precision motion control and is scalable for future industrial applications.

3. Experimental Results

The experimental platform consists of a custom-built 6-DOF robotic arm constructed from a modified 3D printer frame. Each joint is actuated by standard servo motors (MG996R and SG90), with the end-effector equipped with a simple gripper. The control system is based on an Arduino Uno microcontroller, which sends PWM signals to the servos to set joint angles.

Commands are transmitted wirelessly via Bluetooth (HC-06 module) from a custom-developed Android application. The application allows the user to define and save a sequence of target positions (trajectory points), which are then sent to the Arduino to execute in real time. Each experiment was repeated three times under identical operating conditions to evaluate system consistency and repeatability. The same trajectory file and environmental conditions were maintained across all trials.

The system used direct mapping between saved joint positions and PWM signals. Each joint angle in the saved file corresponded to a calibrated PWM value, which the Arduino sent to the servos in timed intervals. This method assumes that each servo reaches its target position reliably without verification. A fixed delay was introduced between each motion command to ensure the servo reached the desired position before executing the next command as shown in Figure 8.

The robot was tasked with following a set of five predefined waypoints in 3D space. These waypoints were saved in the Android application and transmitted to the robotic arm as a sequential list. Upon receiving the data, Arduino executed the motion plan by sending the corresponding PWM values to each joint. This process simulated a basic pick-and-place or part-inspection task.

One of the key features of the system is the ability to record and store motion sequences directly from the Android application. During the training phase, the user can move the robotic arm to desired positions manually through the app interface. These positions are then saved to a local file on the Android device as shown in Figure 9. Once saved, the user can run the file allowing the robotic arm to execute the entire motion sequence autonomously and repeatedly without further user intervention. This functionality enables:

• Offline execution of repetitive tasks (e.g., pick-and-place, inspection routines)

• Looped motion playback, allowing the robot to perform the saved motion indefinitely

• Easy editing or replacement of motion profiles by updating the file in the app

This feature significantly improves the system’s usability for repeated tasks in educational and lightweight industrial environments.

The system's performance was evaluated using the following metrics.

• Task Completion Time: The total time taken to complete the full 5-point trajectory.

• Repeatability: The positional consistency across three repeated executions of the same trajectory.

• Motion Smoothness: Assessed based on the continuity of motion, as well as the absence of servo jitter or stalling.

The joint angles for each of the five trajectory waypoints, based on the saved trajectory file from the Android app, are provided in Table 3.

4. Results

The robotic arm was evaluated in the University of Texas Permian Basin (UTPB) Mechatronics Laboratory. Through three experimental trials, the system consistently executed the predefined trajectory. The mean trajectory completion time was 13.2 s, with a variation of ± 0.3 s. Visual inspection verified that all joint angles reached their expected positions with minimal observable deviation, despite the presence of minor backlash in the 3D-printed mechanical structure.

Notwithstanding these mechanical limitations, the robotic arm successfully returned to its initial position at the conclusion of each operational cycle. The total time required to complete the five-point trajectory is summarized in Figure 10.

To further quantify system performance, the positioning repeatability was evaluated across three trials. The observed deviation at the end-effector was within an estimated range of ±2–3 mm, primarily influenced by servo backlash and mechanical tolerances of the 3D-printed structure. The system demonstrated stable motion behavior without noticeable oscillation or instability. The robotic arm was able to handle lightweight objects (up to approximately 0.9 g) without performance degradation, confirming its suitability for low-payload industrial applications.

5. Conclusion

All experimental evaluations were conducted in the University of Texas Permian Basin (UTPB) Electrical Laboratory. The experimental results demonstrate that the system can reliably traverse predefined waypoints and successfully complete basic pick-and-place tasks with reasonable accuracy. The motion file storage and replay capability allows users to save and repeatedly execute motion sequences, providing an effective solution for automating repetitive operations. The robotic arm exhibited satisfactory performance for lightweight tasks under both static and dynamic conditions, achieving an average trajectory completion time of 13.2s.

Future work will focus on extending the Android application to integrate a high-definition camera system to further improve operational flexibility and ease of use for operators. The proposed robotic arm system represents an accessible and flexible platform for basic motion control tasks in light industrial applications within the United States. The autonomous motion storage and replay functionality significantly enhances system usability by enabling repeated task execution with minimal human intervention. With further refinement, this system has the potential to serve as a foundational platform for more advanced robotic and automation applications.

References