Proposed System Architecture

We don't sort raw garbage. We propose working downstream of civic dry-waste separation, where optical AI can perform reliably.

Circular waste flow concept illustration
AI-generated concept · visual illustration only

The 7-Step Automated Pipeline

01

Waste Input

Unsorted municipal dry waste enters the infeed system.

Nepal context: Designed for Nepal's dry municipal streams

02

AI Camera Inspection

High-speed industrial cameras scan passing objects.

Nepal context: Designed for Nepal's dry municipal streams

03

Sensor Detection

Multispectral & NIR sensors detect material density & reflectance.

Nepal context: Designed for Nepal's dry municipal streams

04

Material Classification

Deep neural networks classify polymers, metals, and cardboard.

Nepal context: Designed for Nepal's dry municipal streams

05

Robotic Arm Pick & Place

Automated actuators and robotic diverters separate target items.

Nepal context: Designed for Nepal's dry municipal streams

06

Sorted Collection Bins

Pure mono-material streams deposited into dedicated baling bins.

Nepal context: Designed for Nepal's dry municipal streams

07

Cloud Analytics

Real-time IoT telemetry logs recovery rates and contamination.

Nepal context: Designed for Nepal's dry municipal streams

This is a proposed system — currently in active research and prototyping.