Overview

A company has harnessed the power of OpenVINO and integrated it into its everyday application, garbage collection. By enhancing machine vision with OpenVINO technology, the traditional two to threeman garbage truck operation can now be performed with just the driver or a driver with an operator standing by. The driver simply drives up to the curb where the garbage can is positioned, and the machine vision analyzes and determines the distance and the width of the grip needed to pick up the garbage can, and the robotic arm moves into action, picking up the garbage can and dumping the garbage into the truck.
This application brings safety and operation efficiency to garbage collection while minimizing hours of man power and heavy labor.
Intel OpenVINO™ is a toolkit designed to optimize and deploy AI inference on edge devices. It enables developers to convert pre-trained models from popular frameworks into a format that runs efficiently on Intel hardware. With features like model optimization, support for various architectures, and tools for real-time analytics, OpenVINO enhances performance for applications like computer vision, natural language processing, and robotics. Its ability to facilitate low-latency processing makes it ideal for edge AI applications, enabling smarter, faster decision-making in environments with limited resources.
Challenges
The challenges faced when deploying a computer into a garbage collection truck are no different to general invehicle deployments.
The computer must withstand a vehicle's constant shock and vibration conditions, and the installation location would most likely be confined with limited ventilation. So, the computer must be able to operate in out-of-norm ambient temperatures; and still provide enough processing power to analyze and execute tasks.
The computer must also provide connectivity for actuators, sensors, and cameras to control the robotic arm for bin pick-up.
Different to general vehicles, the collection truck's embedded robotic arm and industrial-grade compressor create droning/ humming vibration when in operation. This may cause connection issues to cables, installed modules, computer's overall structural integrity, etc. over time.
Solution

A system integrator decided to deploy Neousys POC VTC and utilize Intel® OpenVINO™ for the garbage collection truck project.
E-Mark certified, the ultra-compact POC VTC features Intel's i3-N305 processor and supports Open Visual Inference and Neural Network Optimization (OpenVINO) for deep learning and inference modeling on light AI applications. In this case, OpenVINO can accurately determine the distance the robotic arm needs to reach out to grasp and pick up the garbage bin.
It has four 802.3at PoE+ ports for data and power connection to cameras via a single cable, simplifying connectivity. And to ensure connection reliability in vibration conditions, there is a screw-lock mechanism for every connector.
The POC VTC also comes with dual SocketCAN where it features network-like communication allowing multiple applications to access CAN device simultaneously, bringing function-flexibility to in-vehicle applications.
