Context

RC vehicles are normally controlled by a remote controller via radio signals. Although a good solution for short distancies, it fails in operations that require the operator and vehicle to be separated by a long distance. For example, there could be a need for a RC vehicle accomplish a task where humans shouldn't be directly involved in.

WebRemote is a service that allows the user to control a vehicle through the web. By having access to the Internet, the operator and the vehicle can interact independently of the distance that separates them.

Features

The web page used to control the vehicle allows for two methods of interaction:

1. by clicking in the directional buttons
2. by voice

The web page also displays a video stream from the perspective of the vehicle.

Due to the nature of the technology, remote controlling the car through the web comes with added lag. In order to prevent the vehicle from crashing into walls or any other obstacle, it comes with ultrasound sensors that allow for collision detection.

Visual Concept

Web page:

Remote control car:

Architecture

• Ultrasonic sensor
• In a python file, Raspberry Pi continuously calculates the distance to obstacle and sends it to Kafka.
• Raspberry Pi Camera
• In another python file, Raspberry Pi chooses the resolution and captures a photo then encode the image in bytes and sends to Kafka.
• Sound sensor
• A voice sensor is used to capture the voice commands made by a user.
• Sbrick
• Sbrick receives comands from Raspberry Pi through bluetooth, and makes the motor and servo move or stop.
• Raspberry Pi
• Raspberry has python files running, two already explained before. The next file receives messages from Kafka that can be drive or stop messages and sends comands to sbrick accordingly.
• Kafka cluster
• The kafka broker stores the various information shared between the Raspberry PI, the Springboot application and the webpage. It has the following topics:
  • p1g2drive: for receiving the commands and sending them to the car
  • p1g2ultrasonic: for receiving the current distance of the car from an obstacle and sending it to the web browser
  • p1g2camera: for receiving the camera frames and sending them to the web browser
• SpringBoot Server
• The Springboot web application coordinates the flow of information, have to deal with processing of the ultrasonic sensor, camera sensor and commands data in real time. It also provides REST API for the web page to query information about the camera frames and the ultrasonic sensor data. It connects to the PostgreSQL database to store the history of the commands executed by the user, with the respective timestamp.
• Web Browser
• The web page receives the camera frame data and the ultrasonic sensor data in real time and shows them to the user. It also has a set of buttons to control the car.
• PostgreSQL database
• We use a PostgreSQL database to store the commands executed by the user. Each record has associated with it the timestamo when it was executed. This type of database was chosen in order to easily retrieve information between a certain time interval.

Technical Requirements

In terms of data, the application will handle live streaming of frame data from the camera, ultrasound data and the JSON commands sent by the web page.

• Raspberry Pi 3 Model B V1.2
• Raspberry Pi Camera V2.1
• 2 x HC-SR04 (Ultrasound sensor)
• Waveshare Sound sensor V2 (microphone)
• 1 or 2 breadboard(s)
• Jumpers
• Lego car:
• Sbrick (controller)
• LEGO Extension Wire
• LEGO Power Functions Servo Motor
• LEGO Power Functions L-Motor
• LEGO® Power Functions Battery Box
• Batteries