Smart Sort Arm
Automated Intelligent Sorting Robotic Arm System
2026 Academic Year
Table of Contents
Table of Contents.................................................................................................................................
Abstract...............................................................................................................................................
Introduction.........................................................................................................................................
Background and Motivation..................................................................................................................
Problem Statement...............................................................................................................................
Project Objectives.................................................................................................................................
Mechanical Synthesis.....................................................................................................................
Sensor Integration..........................................................................................................................
Embedded Control..........................................................................................................................
Mechanical Design...............................................................................................................................
Overview of Mechanical Assemblies.....................................................................................................
Conveyor Transport Assembly..............................................................................................................
Robotic Manipulator Base and Product Base........................................................................................
Sorting Storage Units (Collection Bins)..................................................................................................
DC Motor and Conveyor Integration.....................................................................................................
Torque and Speed Regulation.........................................................................................................
Pulse-Width Modulation (PWM) Control........................................................................................
Power Isolation...............................................................................................................................
Sensor Integration...............................................................................................................................
PR18-8DN Inductive Proximity Sensor..................................................................................................
Electrical and Output Integration...................................................................................................
Mechanical and Environmental Robustness.................................................................................
Interfacing PR18-8DN with ESP32.......................................................................................................
Recommended Solution: Pre-built 24 V Opto-isolated Relay Module..........................................
Wiring Requirements....................................................................................................................
Advantages for Smart Sort Arm....................................................................................................
Abstract
This graduation project presents the design and implementation of an automated, intelligent sorting
robotic arm system integrated with a conveyor belt mechanism. The primary objective of the system
is to automate the industrial process of material classification, enhancing efficiency and reducing
human error.
The system utilizes a conveyor belt to transport products through an inspection zone equipped with
inductive and capacitive proximity sensors to identify and differentiate between metal, wood, and
plastic materials. Upon detection, a high-precision robotic arm—actuated by five stepper motors—
picks and places the identified item into one of three dedicated sorting bins based on its material
composition.
The hardware architecture is driven by an ESP32 microcontroller running custom Arduino firmware,
while a dedicated web application (HTML/CSS) provides an intuitive user interface for system
monitoring and control. The mechanical structural bodies were modeled using SolidWorks and
AutoCAD to ensure kinematic stability and structural integrity.
Introduction
Background and Motivation
In modern industrial automation, the separation of heterogeneous material streams—such as
ferrous metals, non-ferrous metals, plastics, and wood—constitutes a fundamental operation in
manufacturing, recycling, and logistics. Manual sorting, while flexible, suffers from low throughput,
high labor costs, and significant variability due to human fatigue and perceptual limitations.
With the advent of Industry 4.0 paradigms, intelligent cyber-physical systems have emerged as the
preferred solution. These systems integrate robotic manipulators, multi-modal sensing, and real-
time control networks to achieve high-speed, high-accuracy sorting with minimal downtime.
The motivation for the present work stems from the growing need for compact, low-cost sorting
cells suitable for small-to-medium enterprises (SMEs) and educational research platforms. Existing
industrial sorters are often prohibitively expensive, bulky, and require specialized programming. By
designing a modular "Smart Sort Arm" that combines a 5-degree-of-freedom (DoF) robotic arm, a
continuously operating conveyor belt, and an inductive proximity sensor (PR18-8DN), this project
3 / 13
aims to demonstrate a scalable proof-of-concept that balances performance, cost, and educational
value.
The use of an ESP32 microcontroller further enables wireless monitoring and remote control,
aligning with current trends toward Internet-of-Things (IoT) integration in automation.
Problem Statement
Industrial sorting environments present three interrelated engineering challenges: material
discrimination, mechanical handling, and real-time coordination.
First, reliable real-time discrimination between metals and non-metals (plastics, wood, ceramics) is
required. Inductive proximity sensors, such as the PR18-8DN, can detect conductive objects without
contact, but their output is a 24 V NPN (open-collector) signal, which is not directly compatible with
the 3.3 V or 5 V logic levels of typical microcontrollers. A robust signal conditioning stage must
therefore be designed to interface the sensor to an ESP32 without introducing latency or noise.
Second, the mechanical subsystem must ensure continuous transport of mixed materials while
allowing precise robotic pick-and-place operations. The conveyor belt must maintain a constant
linear speed, and the robotic arm must achieve sufficient joint accuracy and repeatability to grasp
objects of varying sizes (20-50 mm) without collision. The arm's 5-DoF architecture introduces non-
linear kinematics, requiring careful mechanical synthesis and trajectory planning.
Third, the embedded control system must handle multiple concurrent tasks: reading the sensor
input, driving stepper motors for both the conveyor and the arm, and maintaining a wireless user
interface. Any delay or asynchrony between sensing and actuation can lead to missed or mis-sorted
objects, reducing overall system effectiveness.
Thus, the core problem is to develop an integrated sorting cell that overcomes these electrical,
mechanical, and software synchronization challenges while remaining within a modest budget and
being replicable for academic purposes.
Project Objectives
To address the above problem, this project defines four primary objectives, each with specific
technical sub-targets.
Mechanical Synthesis
To design and fabricate a balanced 5-DoF robotic arm and a modular conveyor belt framework using
CAD software. The arm shall use stepper motors with planetary gearboxes to achieve a positioning
4 / 13
repeatability of +/-1 mm at the end-effector. The conveyor belt shall have an adjustable speed range
of 0.05-0.2 m/s and a sorting zone length of at least 300 mm. All mechanical components shall be
manufacturable via 3D printing or readily available aluminium profiles.
Sensor Integration
To implement a robust signal conditioning chain for the PR18-8DN inductive proximity sensor. This
includes designing a level-shifting circuit (e.g., a pull-up resistor or optocoupler) that converts the 24
V NPN output to a 3.3 V logic signal suitable for the ESP32, while preserving the sensor's 350 Hz
maximum switching frequency. Additionally, the system shall be capable of distinguishing between
ferrous and non-ferrous metals by exploiting differences in the sensor's effective sensing distance.
Embedded Control
To program an ESP32 microcontroller to execute real-time tasks: (i) polling the conditioned sensor
signal with interrupt-driven detection, (ii) generating step and direction pulses for up to six stepper
motors (five for the arm, one for the conveyor), and (iii) hosting a responsive web interface via Wi-Fi
for monitoring and manual override.
5 / 13
Mechanical Design
Overview of Mechanical Assemblies
The mechanical layout of the smart sorting system is split into three main sub-assemblies: the
Conveyor Transport Subsystem, the Robotic Manipulator Base, and the Sorting Storage Units. Each
sub-assembly is designed for modularity, ease of fabrication, and integration with the electronic
control system.
Conveyor Transport Assembly
The conveyor belt serves as the continuous feeding mechanism for the system. Its design
emphasizes rigidity, adjustability, and compatibility with the sensing and actuation modules.
• Support Framework: Fabricated using lightweight, rigid square aluminum/steel profiles to
minimize vibration during operation.
• Drive System: The belt is stretched across tensioned rollers. Underneath the main frame, a
dedicated mounting bracket holds the primary actuator (motor/gearbox assembly) to drive
the belt smoothly.
• Sensor Bracket Zone: Positioned along the track to hold the capacitive and inductive
proximity sensors at an optimal distance from the passing items to guarantee accurate
readings.
6 / 13
Robotic Manipulator Base and Product Base
The sorting arm is the core mechanical actuator of the system. Its design focuses on reach, payload
capacity, and repeatability.
• Structure: Designed using SolidWorks, the arm features a robust structural body optimized
for the weight distribution of the 5 stepper motors and their corresponding joints and
linkages.
• Product Base Plate: Designed via AutoCAD, the center plate acts as the structural foundation
where the arm is anchored. It includes precise CNC/laser-cut mounting holes for securing
the main rotating axis, distributing the dynamic forces generated during high-speed pick-
and-place maneuvers.
• Kinematics: Driven by 5 independent stepper motors, providing the necessary Degrees of
Freedom (DoF) to reach the conveyor terminal point and swing smoothly to the sorting bins.
Sorting Storage Units (Collection Bins)
The tri-box configuration consists of a specialized 3-box sheet metal or acrylic array fixed adjacent to
the robotic arm. The boxes feature open tops and are spaced at exact angular coordinates
corresponding to the pre-programmed destination paths of the robotic arm. This ensures that the
arm can drop materials cleanly into the designated Metal, Wood, or Plastic bins without colliding
with the edges.
7 / 13
DC Motor and Conveyor Integration
While the robotic arm utilizes high-precision stepper motors for exact coordinate positioning, the
continuous transport system relies on a DC Motor. In industrial automation, DC motors are preferred
for conveyor belts due to their high starting torque, smooth continuous rotation, and
straightforward speed regulation.
Torque and Speed Regulation
Unlike the stepper motors which move in discrete angular steps, the DC motor provides continuous
rotational kinetic energy. It is paired with a mechanical gearbox (Geared DC Motor) to lower the
RPM while significantly multiplying the output torque. This ensures the belt can move smoothly even
when loaded with heavier products (like dense wood or metal blocks).
Pulse-Width Modulation (PWM) Control
The speed of the conveyor belt must be carefully synchronized with the sampling rate of the
proximity sensors. By utilizing an H-bridge motor driver (such as an L298N or modern equivalent)
connected to the ESP32 microcontroller, the system applies PWM. Adjusting the PWM duty cycle
changes the average voltage delivered to the motor, allowing precise speed calibration.
Power Isolation
Because DC motors can introduce inductive voltage spikes (back EMF) into the circuit during startup
or sudden stops, its power rail is completely isolated from the logic power rail of the ESP32 using
optocouplers or independent power steps to prevent microcontroller resets.
8 / 13
Table 1: DC Motor Specifications for Conveyor System
Parameter Value
Motor Type Brushed/Geared DC Motor
Voltage Range 12-24 V DC
Speed Control PWM via H-Bridge (L298N)
Belt Speed Range 0.05 - 0.20 m/s
Sorting Zone Length >= 300 mm
Power Isolation Optocoupler / Independent Rails
Sensor Integration
PR18-8DN Inductive Proximity Sensor
The PR18-8DN inductive proximity sensor is the ideal component for a metal sorting robot. Its high-
speed response and robust build make it a reliable choice for accurate, non-contact metal detection
in automated systems. The sensor generates an electromagnetic field to detect metal without
physical contact.
Table 2: PR18-8DN Sensor Specifications
Parameter Specification
Type Inductive Proximity Sensor
Model PR18-8DN (Autonics)
Sensing Distance 8 mm (+/- 10%)
Response Frequency 350 Hz
9 / 13
Parameter Specification
Output Type NPN, Normally Open (NO)
Supply Voltage 12-24 VDC (10-30 VDC range)
Control Output <= 200 mA
Protection Rating IP67
Operating Temperature -25 to 70 degrees C
Housing M18 threaded nickel-plated brass
Weight Approx. 110 g
Electrical and Output Integration
• Output Type: NPN, Normally Open (NO). In this configuration, the output connects to 0V
(GND) when metal is detected.
• Wiring: Simple 3-wire connection (Brown: +V, Blue: 0V, Black: Output).
• Supply Voltage: 12-24 VDC (operating range: 10-30 VDC), allowing for flexible power supply
integration.
• Control Output: Less than or equal to 200 mA.
• Protection Circuits: Built-in protection against surges, short circuits, and reverse polarity
ensures high reliability in an industrial environment.
Mechanical and Environmental Robustness
• Protection Rating: IP67, providing excellent resistance to dust and temporary immersion in
water.
• Ambient Temperature: Operates from -25 to 70 degrees C (storage: -30 to 80 degrees C).
• Vibration Resistance: Withstands 1.0 mm amplitude at 10 to 55 Hz.
• Connection: 2-meter standard PVC cable.
• Weight: Approximately 110 g.
Interfacing PR18-8DN with ESP32
The PR18-8DN is an industrial inductive proximity sensor that operates from a 12-24 V DC supply and
provides an NPN (open-collector) output. In contrast, the ESP32 microcontroller accepts logic signals
10 / 13
with a maximum voltage of 3.3 V. Directly connecting the sensor's 24 V output to an ESP32 input pin
risks permanent damage to the microcontroller.
Furthermore, the sensor's NPN output cannot directly drive a relay coil because the coil's inductive
kickback and current demand exceed the sensor's safe output rating. Therefore, a signal conditioning
stage is required to: (i) level-shift the 24 V logic to 3.3 V logic, (ii) provide electrical isolation if
necessary, and (iii) allow the ESP32 to control higher-power actuators through a relay.
Recommended Solution: Pre-built 24 V Opto-isolated Relay Module
The most practical and safe solution is a 1-channel 24 V relay module with opto-isolation. Such a
module contains:
• A mechanical relay (typically rated for 10 A at 30 V DC or 250 V AC).
• An optocoupler that isolates the ESP32's 3.3 V logic side from the 24 V relay coil side.
• A transistor driver circuit and a flyback diode across the relay coil.
Wiring Requirements
• The module's VCC pin must be connected to a separate 24 V DC power supply (the same
supply that powers the PR18-8DN sensor). Do not connect VCC to the ESP32's 3.3 V pin.
• The module's GND pin is connected to both the negative terminal of the 24 V supply and the
ESP32's GND. This common ground is essential for the control signal.
• The IN (control) pin is connected to an ESP32 digital output pin. When the ESP32 sets this pin
to HIGH (3.3 V), the optocoupler activates, turning on the relay.
Advantages for SmartSortArm
• Complete galvanic isolation between the ESP32 and the high-power actuator circuit.
• Simple to implement with minimal additional components.
• Robust against electrical noise from the sorting mechanism.
11 / 13
Table 3: Sensor-to-Microcontroller Interface Parameters
Parameter Value
Sensor Output 24 V NPN (Open-Collector)
Microcontroller Input 3.3 V Logic (ESP32)
Level-Shifting Method Opto-isolated Relay Module
Relay Rating 10 A at 30 V DC / 250 V AC
Isolation Type Galvanic (Optocoupler)
Common Ground Required (Sensor GND + ESP32 GND)
12 / 13
13 / 13