From what to build to how to build it — our platform helps you understand, plan, and solve SIH 2026 problem statements with complete insights into requirements, solution approaches, technology stacks, architecture, AI/ML, and implementation roadmaps.
Search for your problem statement, and start building your solution.
| Description: | • Description Healthcare facilities generate large volumes of biomedical waste that require safe, compliant, and efficient handling. Manual collection and segregation increase the risk of contamination, operational inefficiencies, and regulatory challenges. Design and develop an AI-powered, battery-electric autonomous mobile system that automates the collection, identification, segregation, and digital tracking of biomedical waste across hospitals. The solution should leverage AI-enabled vision systems to classify waste, intelligently segregate it into designated compartments, and provide end-to-end traceability while minimizing human exposure to hazardous materials and improving safety, operational efficiency, and regulatory compliance. Using Autodesk Fusion, students must demonstrate a complete product development lifecycle-from concept ideation to manufacturing-ready product, delivering an innovative, scalable, and solution for next-generation healthcare waste management. • Participation Guidelines For Idea Submission: • Each student team will submit a PowerPoint presentation (5-7 Slides) with conceptual sketches, research, and relevant images. • NO design files are required at this stage. The actual design must be created ONLY during the Grand Finale. • Designs should be created using ONLY Autodesk Fusion and not copied or taken from any other source. • AI Generated content is NOT ALLOWED. For Grand Finale: Students must use Autodesk Fusion within the given time period and present: • PPT explaining the final project • Public link of the fully developed Autodesk Fusion design model • Use of Generative Design for optimization will be an added advantage. • Simulation and Analysis will be advantageous • Motion study and exploded assembly view • Hi-res rendered images • Design should be capable of developing a prototype with a focus on Cost, Manufacturability, Scalability, and Quality. Note: Teams coming with pre-designed files will be disqualified. • Attach Marking Criteria Table here* • Faculty (SIH SPOC) Form Teams choosing to submit idea for Autodesk’s problem statement are required to request their faculty (SIH SPOC) to fill this 'Mandatory Form'. • Autodesk Fusion • Autodesk Fusion is a cloud-based 3D modeling, CAD, CAM, CAE, and PCB software platform for professional product design and manufacturing. • Students and educators can click here to get FREE access to Fusion. |
| Organization: | Autodesk |
| Department: | Autodesk Education Experience |
| Category: | Software |
| Theme: | MedTech / BioTech / HealthTechtd> |
| Youtube Link: | Not provided |
| Dataset Link: | Additional Information regarding PS https://drive.google.com/drive/folders/1KYkkQjRbpq--YRF5sgVndjLGgItqLr1b |
| Contact info: | Not provided |
Healthcare facilities generate substantial quantities of hazardous biomedical waste, where traditional manual handling introduces severe contamination risks, operational bottlenecks, and compliance violations [PS]. To address this critical challenge, this dossier defines a complete engineering strategy for designing and building an AI-powered, battery-electric autonomous mobile robot (AMR) system integrated with automated waste identification, multi-compartment segregation, and real-time cloud-based digital tracking [PS]. Built strictly in compliance with the Bio-Medical Waste Management Rules (2016) [RESEARCHED] and executed using Autodesk Fusion for the full product development lifecycle [PS], the proposed system bridges advanced mechanical product design, edge vision AI, autonomous indoor navigation, and digital chain-of-custody tracking. This document serves as an exhaustive blueprint enabling a 4–6 student hackathon team to execute the concept during the Smart India Hackathon (SIH) timeline, providing clear boundaries between the hackathon prototype and production deployment.
| Attribute | Details | Evidence Tag |
|---|---|---|
| Problem Statement Number | SIH26115 | [PS] |
| Title | Design and Develop a Smart Mobile Medical-Waste Collection and Segregation System | [PS] |
| Organization | Autodesk | [PS] |
| Department | Autodesk Education Experience | [PS] |
| Category | Software / Mechanical Hardware Hybrid | [PS] |
| Theme | MedTech / BioTech / HealthTech | [PS] |
| Submission Deadline | 30 September 2026 | [PS] |
| Ideas Counter | 0 / 500 | [PS] |
| Official Dataset Link | Google Drive Folder (1KYkkQjRbpq--YRF5sgVndjLGgItqLr1b) |
[PS] |
To eliminate human contact with hazardous hospital waste by engineering a battery-electric autonomous mobile robot designed in Autodesk Fusion that identifies biomedical waste categories using edge computer vision, mechanically segregates items into regulatory-compliant compartments, and maintains tamper-proof digital chain-of-custody tracking across hospital wards [INFERENCE].
The problem statement has been broken down into atomic, explicit, and derived requirements. All explicit conditions from the official Autodesk problem description are strictly preserved without modification [PS].
| Stakeholder | Goal | Current Pain | System Need |
|---|---|---|---|
| Sanitation / Utility Workers | Transport biomedical waste safely without exposure to pathogens or needle-stick injuries [KNOWN]. | High physical risk of infection, accidental sharp punctures, and lifting heavy ergonomic loads [KNOWN]. | Touchless waste deposit mechanism, automated transport, and sealed puncture-proof compartments [INFERENCE]. |
| Ward Nurses & Doctors | Dispose of clinical waste quickly at point-of-care without leaving patient care zones [KNOWN]. | Overflowing waste bins in wards; confusion over complex color-coding sorting rules causing missegregation [KNOWN]. | Automated mobile collection vehicle calling at scheduled intervals with AI assistance for waste categorization [INFERENCE]. |
| Hospital Infection Control Officer (ICO) | Eliminate hospital-acquired infections (HAIs) and maintain strict cross-contamination barriers [KNOWN]. | Manual waste carts moving through public corridors spreading bio-aerosols and surface contamination [KNOWN]. | Sealed, UV-C sanitized, autonomous waste transport following isolated logistics routes [INFERENCE]. |
| Hospital Waste Compliance Manager | Ensure 100% adherence to Central Pollution Control Board (CPCB) Bio-Medical Waste Management Rules 2016 [RESEARCHED]. | Fines and legal exposure due to inaccurate manual logging, missing weight records, and bag misplacement [KNOWN]. | Automated digital logging (time, weight, category, location) synced directly to CPCB/hospital portal [INFERENCE]. |
| Facility Engineering & IT Team | Maintain fleet uptime, monitor robot health, and integrate robotics with hospital logistics [KNOWN]. | Proprietary closed systems that fail to integrate with hospital Wi-Fi, elevators, or floor layouts [KNOWN]. | Open ROS2 architecture, standard REST/MQTT APIs, and Autodesk Fusion open CAD models for custom spare parts [INFERENCE]. |
| Proposed System Feature | Justification Source Requirement | Exact Functionality |
|---|---|---|
| 4-Compartment Color-Coded Segregation Vault | Requirement 1 & 2 [PS], BMW Rules 2016 [RESEARCHED] | Bins dedicated to Yellow (soiled/pathological), Red (contaminated plastics), Blue (glassware), and White (sharps box) with motor-driven locking lids [INFERENCE]. |
| Edge Vision AI Camera with High-Brightness LED Ring | Requirement 2 [PS] | Live capture of presented waste, inferring material class using lightweight CNN model, displaying bin prompt on screen [INFERENCE]. |
| Generative Design Structural Chassis | Requirement 4 & 5 [PS] | Autodesk Fusion Generative Design workspace used to minimize chassis mass while maintaining load capacity for 50 kg waste payload [INFERENCE]. |
| LiDAR + Ultrasonic Obstacle Avoidance Suite | Requirement 1 [PS] | Dynamic path planning through crowded hospital corridors, stopping smoothly for pedestrians, wheelchairs, and stretchers [KNOWN]. |
| Digital Chain-of-Custody Telemetry Module | Requirement 3 [PS] | On-board load cell weight sensor + RFID scanner logging waste origin, bag ID, timestamp, and net weight to MQTT cloud server [INFERENCE]. |
| Motion Study & Exploded View CAD Output | Requirement 5 [PS] | Autodesk Fusion Animation workspace producing step-by-step mechanical assembly and functional compartment opening video [INFERENCE]. |
| UV-C Internal Sterilization Module | Requirement 3 [PS] (Safety enhancement) | Internal UV-C LED strips activating between ward visits to sanitize compartment inner walls and eliminate bio-aerosols [INFERENCE]. |
.f3d, .step, or .igs files will be disqualified [PS]. All CAD geometry must be generated on-site within Autodesk Fusion during the hackathon timeline [PS].| Data Type | Purpose | Ideal Real-World Source | Hackathon Substitute | Access Method | Availability | Privacy Sensitivity | Confidence |
|---|---|---|---|---|---|---|---|
| Biomedical Waste Image Dataset | Train AI vision model to recognize gloves, syringes, IV bottles, glass vials, cotton swabs. | Kaggle Medical Waste Datasets / Custom Hospital Annotations | Custom dataset generated by photographing medical props + Kaggle public waste images | Local Download / Roboflow API | High | Low (No Patient Data) | [RESEARCHED] |
| Hospital Indoor Map / Costmap | Autonomous SLAM navigation and route planning. | Real-time LiDAR scan of hospital floor | Simulated 2D floorplan generated in Gazebo / RViz2 | ROS2 Nav2 / Cartographer | High | Medium (Facility Layout) | [INFERENCE] |
| Waste Transaction Logs | Digital chain-of-custody logging (weight, timestamp, bin ID, operator). | Hospital Information System (HIS) API | Mock JSON payload generated by robot micro-controller | MQTT / REST over Wi-Fi | High | Low (System Logs) | [INFERENCE] |
| Autodesk Fusion CAD Parameters | Parametric dimensions for rapid live CAD reconstruction during finale. | Dimensional spec sheets of standard components (motors, battery, camera) | Prerecorded dimension parameter table (.csv / text sheet) |
Manual Entry into Fusion Parameters | High | None | [ASSUMPTION] |
| Official PS Supplemental Data | Verification of exact guidelines provided by Autodesk organizers. | Problem Statement Drive Link [PS] | Extracted instructions from Drive folder | HTTPS Browser Download | High | Public | [PS] |
Approach 2: Fully Autonomous AMR with Edge AI Vision & Multi-Bin Smart Segregation Chute.
+-----------------------------------------------------------------------------------+
| RECOMMENDED SYSTEM OVERVIEW |
| |
| [ Hospital Ward ] |
| │ |
| ▼ |
| +------------------+ Vision AI +-------------------------------------+ |
| | Waste Presented | ─────────────────► | Camera + Edge TPU / Jetson Nano | |
| +------------------+ +-------------------------------------+ |
| │ |
| ▼ |
| +-----------------------------------------------------------------------------+ |
| | Autodesk Fusion Designed Robot Compartment | |
| | +----------------+ +----------------+ +----------------+ +--------------+ | |
| | | YELLOW BIN | | RED BIN | | BLUE BIN | | WHITE BIN | | |
| | | (Soiled/Anat.) | | (Plastics) | | (Glassware) | | (Sharps) | | |
| | +----------------+ +----------------+ +----------------+ +--------------+ | |
| +-----------------------------------------------------------------------------+ |
| │ |
| ▼ |
| +-------------------------------------+ |
| | Central Cloud Chain-of-Custody Log | |
| +-------------------------------------+ |
+-----------------------------------------------------------------------------------+
+-----------------------------------------------------------------------------------+
| SYSTEM ARCHITECTURE |
+-----------------------------------------------------------------------------------+
| 1. PERCEPTION & SENSOR LAYER |
| ├── 2D LiDAR (RPLiDAR A2M8) ──────────── Navigation & SLAM |
| ├── Wide-Angle USB/CSI Camera ─────────── Waste Vision Classification |
| ├── Ultrasonic Sensors (x4) ───────────── Bumper & Cliff Safety |
| └── Load Cell Strain Gauges (x4) ──────── Compartment Weight Tracking |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| 2. EDGE EMBEDDED CONTROL & AI LAYER |
| ├── Main Compute: NVIDIA Jetson Orin Nano / Raspberry Pi 5 + AI Kit |
| │ ├── ROS2 Humble (Nav2, AMCL, Slam-Toolbox) |
| │ └── TensorRT / ONNX Inference Engine (YOLOv8 Waste Classifier) |
| └── Low-Level Controller: ESP32 / STM32 Microcontroller |
| ├── Differential Drive PWM Motor Controllers |
| ├── Servo Actuators for Bin Diverter Lids |
| └── Load Cell HX711 ADC Interface |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| 3. AUTODESK FUSION MECHANICAL CHASSIS LAYER |
| ├── Generative Designed Aluminium Support Framework |
| ├── High-Density Polyethylene (HDPE) Modular Waste Bins (IP54 Sealed) |
| ├── Servo-Driven Rotary Intake Funnel Chute |
| └── Quick-Release Battery Tray with Locking Latch |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| 4. DIGITAL CONNECTIVITY & TELEMETRY LAYER |
| ├── Wi-Fi / LTE Cellular Telemetry Module |
| ├── MQTT Protocol Broker (EMQX) for Real-Time Event Streaming |
| └── Local SQLite Offline Backup Buffer |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| 5. HOSPITAL MANAGEMENT DASHBOARD LAYER |
| ├── React.js / Tailwind CSS Web Frontend |
| ├── Node.js Express REST API Server |
| └── PostgreSQL Database (Waste Metrics, Compliance Audit Logs, CPCB Reports) |
+-----------------------------------------------------------------------------------+
plastic_syringe -> RED_BIN, confidence 0.94) [INFERENCE].[Robot_ID, Ward_ID, Timestamp, Category: RED, Weight: 45g, AI_Confidence: 0.94] into an MQTT payload and transmits it to the hospital cloud database [INFERENCE].| Layer | Selected Technology | Why Chosen | Alternative Considered | Why Alternative Rejected |
|---|---|---|---|---|
| Mechanical CAD / CAE | Autodesk Fusion | Mandatory requirement [PS]. Integrated parametric CAD, Generative Design, FEA simulation, motion study, and cloud rendering [PS]. | SolidWorks / CATIA | Expressly forbidden by problem guidelines; will result in zero score [PS]. |
| AMR Navigation OS | ROS2 Humble Hawksbill | Industry standard for mobile robotics; robust SLAM, Nav2 path planning, modular architecture [KNOWN]. | Custom Python Navigation | Lacks production reliability, costmap management, and community-tested SLAM algorithms [INFERENCE]. |
| Edge Compute Hardware | NVIDIA Jetson Orin Nano / Raspberry Pi 5 + AI Accelerator | High CUDA/NPU performance for real-time vision model inference at low power consumption [KNOWN]. | Standard Laptop on Chassis | Bulky, non-embedded, excessive power drain, unsuitable for real AMR deployment [INFERENCE]. |
| Vision AI Framework | PyTorch / YOLOv8 (ONNX/TensorRT Exported) | SOTA object detection speed and accuracy; lightweight model exportable to edge NPUs [KNOWN]. | OpenCV Color Thresholding | Rule-based color/shape matching fails completely on transparent syringes, bloodied cotton, or varying light [INFERENCE]. |
| Low-Level Microcontroller | ESP32-S3 | Dual-core 240MHz processor, hardware PWM for motors, Wi-Fi/BLE built-in, low cost [KNOWN]. | Arduino Uno | Insufficient RAM, lacks native Wi-Fi/MQTT, single-threaded execution bottleneck [INFERENCE]. |
| Backend & Cloud Telemetry | Node.js + MQTT (EMQX) + PostgreSQL | Asynchronous high-throughput event processing ideal for telemetry streams and CPCB compliance logging [KNOWN]. | Monolithic Python Django | Slower WebSockets telemetry handling for real-time robot fleet tracking [INFERENCE]. |
| Web Dashboard Frontend | React.js + Tailwind CSS + Leaflet/Canvas Map | Fast, responsive UI capable of displaying live robot positions on floorplan overlays [KNOWN]. | Plain HTML/JS | Lacks component state management required for dynamic fleet telemetry dashboards [INFERENCE]. |
Yellow, Red, Blue, White), bounding box, confidence score ($0.00 - 1.00$) [INFERENCE].cmd_vel: $v_x, \omega_z$) to motor drivers [KNOWN].-- Waste Collections Table
CREATE TABLE waste_transactions (
transaction_id VARCHAR(36) PRIMARY KEY,
robot_id VARCHAR(20) NOT NULL,
ward_id VARCHAR(20) NOT NULL,
category VARCHAR(10) CHECK (category IN ('YELLOW', 'RED', 'BLUE', 'WHITE')),
weight_grams REAL NOT NULL,
ai_confidence REAL NOT NULL,
timestamp TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
sync_status INTEGER DEFAULT 0 -- 0: Pending, 1: Synced
);
-- Robot Fleet State Table
CREATE TABLE robot_status (
robot_id VARCHAR(20) PRIMARY KEY,
battery_percentage INTEGER NOT NULL,
current_location VARCHAR(50),
active_state VARCHAR(20), -- 'NAVIGATING', 'COLLECTING', 'CHARGING', 'ERROR'
last_heartbeat TIMESTAMP WITH TIME ZONE
);
To ensure a flawless presentation before SIH judges without relying on unavailable hospital infrastructure:
+-----------------------------------------------------------------------------------+
| LIVE DEMO SIMULATION FLOW |
+-----------------------------------------------------------------------------------+
| STEP 1: CAD & ENGINEERING SHOWCASE (Autodesk Fusion) |
| • Present live parametric Fusion model link, Generative Design iteration tree, |
| FEA stress distribution heatmaps, exploded assembly view video. |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| STEP 2: LIVE AI VISION & MECHANICAL SEGREGATION |
| • Place sample item (e.g., plastic syringe prop) in front of camera. |
| • AI instantly classifies item as "RED BIN - Contaminated Plastic" (Conf: 96%). |
| • Servo motor rotates diverter mechanism and opens RED compartment door live. |
+-----------------------------------------------------------------------------------+
│
▼
+-----------------------------------------------------------------------------------+
| STEP 3: DIGITAL TELEMETRY & DASHBOARD LOGGING |
| • Load cell registers simulated weight (+30g). |
| • Web Dashboard updates instantaneously, displaying updated Red Bin capacity |
| chart and CPCB-compliant audit log entry. |
+-----------------------------------------------------------------------------------+
2026 ROADMAP TO SIH FINALE (Deadline: 30 September 2026)
[Phase 1: Research & Parametric CAD Prep] -> Sep 05 - Sep 12, 2026
├── Define Fusion user parameters table (.csv)
└── Train YOLOv8 model on custom medical waste dataset
[Phase 2: Software MVP & Telemetry] -> Sep 13 - Sep 20, 2026
├── Develop ESP32 microcontroller motor/servo firmware
└── Build React + Node.js cloud telemetry dashboard
[Phase 3: Integration & Fusion Practice] -> Sep 21 - Sep 27, 2026
├── Practice 0-to-1 live Fusion CAD reconstruction (Speed Drills)
└── Conduct end-to-end vision-to-actuation hardware testing
[Phase 4: Final Polishing & Submission] -> Sep 28 - Sep 30, 2026
├── Prepare 5-7 slide PPT presentation (strictly NO AI content) [PS]
└── Execute final live simulation test run
| Test Category | Target Metric / Test Case | Validation Method |
|---|---|---|
| CAD Structural Integrity | Yield Strength Safety Factor $> 2.5$ under 50 kg load | Autodesk Fusion FEA Static Stress Study [PS]. |
| AI Classification Accuracy | $> 90%$ Precision across 4 waste categories | Test dataset evaluation of 200 unseen medical prop images [INFERENCE]. |
| Segregation Speed | $< 3.0\text{ seconds}$ total cycle time (Inference to Door Open) | Stopwatch timing during live hardware test runs [INFERENCE]. |
| Offline Resilience | Zero data loss during 15-minute simulated network drop | Disconnect Wi-Fi during 20 waste deposits; verify auto-sync upon reconnect [INFERENCE]. |
| Obstacle Avoidance | Stop safely within $20\text{ cm}$ of sudden obstacle | Dynamic obstacle pathing test in ROS2 Nav2 environment [INFERENCE]. |
| Risk | Likelihood | Impact | Severity | Mitigation Strategy |
|---|---|---|---|---|
| Disqualification for Pre-made CAD | High | Fatal | High | Practice building full parametric model live inside Fusion during finale from pre-planned parameter sheets [PS]. |
| Misclassification of Hazardous Waste | Medium | Critical | High | Implement strict confidence threshold ($< 0.70$ prompts manual override on UI); default to Yellow (Infectious) if uncertain [INFERENCE]. |
| Hardware Failure / Servo Jamming | Medium | High | High | Use high-torque metal-geared digital servos; add current-sensing stall detection in MCU firmware [INFERENCE]. |
| Wi-Fi Loss in Hospital Basements | High | Low | Low | Local SQLite event buffering on onboard storage ensures zero lost compliance records [INFERENCE]. |
| Disqualification for AI Presentation | Medium | Critical | High | Manually draft all slide content and graphics; strictly avoid AI text/image generators [PS]. |
[ASSUMPTION] The hackathon jury will evaluate the software MVP using a tabletop demonstration setup alongside live Autodesk Fusion CAD models.[ASSUMPTION] The team has access to a standard 3D printer or acrylic laser cutter prior to the finale for prototyping non-CAD hardware elements.[NEEDS VALIDATION] Exact API specifications for interfacing with commercial hospital elevators (varies by elevator vendor).[NEEDS VALIDATION] Specific hospital floor layout dwg/dxf files for creating realistic ROS2 Gazebo simulation worlds.Build a differential-drive Autonomous Mobile Robot featuring an edge AI vision chute that automatically sorts medical waste into 4 internal compartments, designed strictly within Autodesk Fusion and connected to a real-time compliance cloud dashboard [INFERENCE].
Mastery of rapid parametric live CAD modeling in Autodesk Fusion to reconstruct the entire robot assembly from scratch on-site during the Grand Finale without importing forbidden pre-designed files [PS].
Disqualification due to importing pre-existing CAD files or incorporating AI-generated content into presentation slide submissions [PS].
Ensuring fast ($< 150\text{ ms}$), accurate edge vision classification under unpredictable hospital lighting conditions and varied waste deformations [INFERENCE].
1KYkkQjRbpq--YRF5sgVndjLGgItqLr1b) for official organizers' updates [PS]..csv) for motor mounts, bin volumes, and chassis length in Autodesk Fusion [INFERENCE].88 / 100 (High buildability for a combined software/hardware hackathon team utilizing simulated hardware components alongside live Fusion modeling) [INFERENCE].
HIGH CONFIDENCE [INFERENCE] - The problem requirements are completely explicit, regulatory frameworks (BMW Rules 2016) are well-documented, and the technical architecture relies entirely on proven open-source robotics and AI frameworks [KNOWN].
Title: Bio-Medical Waste Management Rules, 2016 - Guidelines & Schedule II
https://en.vikaspedia.in/viewcontent/energy/environment/waste-management/bio-medical-waste-management/bio-medical-waste-management-rules Title: Autodesk Fusion Generative Design & Manufacturing Workflows
https://www.autodesk.com/solutions/generative-design/manufacturing Title: Autonomous Mobile Robots (AMRs) in Hospital Logistics
https://mobile-industrial-robots.com/it/industrie/hospitals What we think about this platform, and how it can help people solve SIH problems.
" 30M+ Google Gemini tokens went into researching, analyzing, and understanding all 233 SIH problem statements to make them easier to explore and solve. "
" The platform makes SIH problem statements much easier to understand. Having everything from the problem details to the procedure in one place. "
" Oh, you got ads? Well, we need to make money too! But, you can use our platform to help you win SIH and take home the prize money. "
Get in touch if you need help bringing your ideas to life, or simply have an interesting idea you'd like to talk through. I'm always open to a conversation and exploring new possibilities together.