Implementation of Smooth Pursuit Behavior for a Humanoid Robot

The main focus of this project is to add smooth pursuit behavior to the humanoid robot. Adding this ability to the robotic head will increase its flexibility for performing vision based oculomotor tasks.

2025-06-28 16:33:04 - Adil Khan

Project Title

Implementation of Smooth Pursuit Behavior for a Humanoid Robot

Project Area of Specialization RoboticsProject Summary

The main focus of this project is to add smooth pursuit behavior to the humanoid robot. Adding this ability to the robotic head will increase its flexibility for performing vision based oculomotor tasks.

Project Objectives

1) Development of Eyes-Head Coordination for smooth Pursuit

  .  Position Control of servo motors

  .  Multi-camera Board integration with RaspberryPi

  .  Converting Image into HSV model

  .  Color Detection

  .  Object Tracking

  .  Pan-tilt movement of eyes

  .  Coordination of eyes 

  .  Pan-tilt-swing movement of neck

   .  Eye Head Coordination

2) Improvement in Mechanical structure for eyes and neck smooth pursuit.

Project Implementation Method

1)  Headless connection 

To connect a Raspberry Pi to a laptop display, we simply used WiFi .

2) Setting up Raspberry Pi                             

Before moving to connect Raspberry Pi to laptop display, we needed an SD card with the OS to install Raspbian on a blank SD card. Connected Raspberry Pi to the laptop via WiFi. Downloaded VNC client and installed it.

3) Position control of servo motor:

Servo control is achieved by sending a servo a PWM signal, a series of repeating pulses of variable width. In modern RC servos the angle of mechanical rotation is determined by the width of an electrical pulse that is applied to the control wire. . For fully left position:

To provide a 1.0ms pulse width, a duty cycle of 5 must be applied. .

For 90° position:

Pulse = 1.5ms , Duty cycle = 7.5     .

For 180° position:                                                  

To provide a 2.0ms pulse width, a duty cycle of 10 must be applied.

Position control by providing the angle to: 

         Dutycycle =  Implementation of Smooth Pursuit Behavior for a Humanoid Robot _1582920350.pngImplementation of Smooth Pursuit Behavior for a Humanoid Robot _1582920351.png  *(desired angle)+10

4) Color Detection

On a computer, color can be represented in many formats. We used HSV model. H= Hue, S = Saturation, and V=Value. The color is detected by using HSV masking.

5) Multiple Cameras Integration

Raspberry pi camera module is designed for connection of two or more cameras on single CSI port on Raspberry Pi.DIP swithes are used for the pin configuration. We connected two Pi cameras with a multiple camera board at camera ports A and C.

6) Object Tracking

In color detection we found the center of the object and its coordinates. When object moves , the center of objects changes. Compared the centers found and proceed the line in the same way. Gave commands to the respective motors accordingly. 

Benefits of the Project

Humanoid robots can perform many humans like activities. They can interact with their environment and with humans. These robots not only interact with environment but can also execute many industrial heavy-duty tasks as well. Humanoid robots like human in their behavioral algorithms and physics can help in medical professionals, scientists, and engineers to deepen the understanding of human conduct in various situations and environments.As human head is the main and most major part of human body. So, all the functions of the body are controlled from the head (brain) through sensory neurons. To perform any type of actions, signals are sent from brain to the specific part of body, and signals of response are also received in head. So, it is very important to focus on main parts of anthropomorphic robots to increase its efficiency with low cast and less weight. In this regard anthropomorphic robot head is main part of robot through which whole body can be controlled. Many animals have almost all their abilities in their heads.

Technical Details of Final Deliverable

Raspberry Pi 3 Model B

Pi Cameras Rev 1.3 (5 MP)

Servo Motors(S3003)

Servo Motors (HS485BB)

Final Deliverable of the Project Hardware SystemType of Industry Education Technologies Artificial Intelligence(AI), RoboticsSustainable Development Goals Quality EducationRequired Resources
Item Name Type No. of Units Per Unit Cost (in Rs) Total (in Rs)
Total in (Rs) 55950
Raspberry Pi Equipment160006000
Memory Cards (32 GB, 16GB) Equipment210002000
Raspberry Pi casing Equipment1450450
Pi Cameras Equipment332009600
Camera Strips Equipment3300900
Statue Equipment120002000
Mechanical Structure (Eyes) Equipment140004000
Mechanical Structure (Neck) Equipment130003000
Servo Motors (HS485HB) Equipment3450013500
Servo Motors (S3003) Equipment412004800
Battery Equipment112001200
Connecting wires, Breadboard Equipment1500500
Mask Equipment115001500
Transport Miscellaneous 140004000
Experimental motors, pan-tilt structure Miscellaneous 125002500

More Posts