Edge Impulse TinyML Final Year Projects 2026
Edge Impulse is a leading platform for collecting data, training TinyML models and deploying to microcontrollers. Projects typically use Studio DSP blocks, FOMO or classification learning blocks, and export firmware for Arduino, ESP32, STM32 or Nordic boards.
Report accuracy, latency, RAM and flash usage. Datasets may be public (Speech Commands, ESC-50, HAR) or custom-collected.
Tools & Platforms
Edge Impulse Studio
FOMO
Arduino
ESP32
STM32
EON Tuner
Best Edge Impulse TinyML Project Topics 2026
55+ topics with tools and datasets used.
| # | Edge Impulse TinyML Project Topic | Tools / Datasets |
|---|---|---|
| 🎤 Audio · Keyword Spotting · Sound Classification | ||
| 01 | AudioKeyword Spotting (Yes/No / Custom Wake Word) with Edge Impulse | Edge Impulse · Google Speech Commands |
| 02 | AudioEnvironmental Sound Classification (Urban / Indoor Events) | Edge Impulse · ESC-50 subset |
| 03 | AudioCough / Sneeze / Snore Detection for Health Wearable | Edge Impulse · custom audio dataset |
| 04 | AudioGlass-Break / Alarm Sound Detection Security Node | Edge Impulse · synthetic + field audio |
| 05 | AudioBaby Cry Detection and Parent Alert Pipeline | Edge Impulse · public cry datasets |
| 06 | AudioMulti-Keyword Spotting with Continuous Inferencing | Edge Impulse · DSP blocks · MCU |
| 07 | AudioVoice Activity Detection (VAD) before KWS Stage | Edge Impulse · energy + ML features |
| 08 | AudioIndustrial Machine Sound Anomaly from Microphone | Edge Impulse · factory audio samples |
| 📷 Vision · Image Classification · FOMO · Object Detection | ||
| 09 | VisionPerson / No-Person Detection with FOMO on MCU Camera | Edge Impulse FOMO · OV7670 / HM01B0 |
| 10 | VisionHand Gesture Classification from Low-Resolution Frames | Edge Impulse · custom gesture images |
| 11 | VisionDigit / Symbol Recognition for Tiny Display Capture | Edge Impulse · MNIST-style set |
| 12 | VisionPlant Disease Leaf Classification on Edge Device | Edge Impulse · PlantVillage subset |
| 13 | VisionFace Mask / PPE Detection with Quantized Model | Edge Impulse · custom PPE images |
| 14 | VisionObject Counting on Conveyor with FOMO | Edge Impulse FOMO · industrial images |
| 15 | VisionVisual Wake Words Benchmark on Constrained Hardware | Edge Impulse · VWW dataset |
| 16 | VisionProduct SKU Classification for Retail Edge Camera | Edge Impulse · custom product set |
| 📐 IMU · Gesture · Motion · Activity | ||
| 17 | IMUIMU Gesture Recognition (Swipe, Circle, Shake) | Edge Impulse · Arduino Nano 33 IMU |
| 18 | IMUHuman Activity Recognition (Walk, Run, Sit, Fall) | Edge Impulse · HAR public datasets |
| 19 | IMUFall Detection System for Elderly Care Wearable | Edge Impulse · fall simulation data |
| 20 | IMUExercise Repetition Counting from Wrist IMU | Edge Impulse · workout IMU logs |
| 21 | IMUAir-Mouse / Cursor Control via Hand Motion | Edge Impulse · continuous motion data |
| 22 | IMUMachine Vibration Pattern Classification | Edge Impulse · accelerometer logs |
| 23 | IMUSports Move Classification (Swing, Kick, Throw) | Edge Impulse · sports IMU dataset |
| ⚠️ Anomaly Detection · Predictive Maintenance | ||
| 24 | AnomalyMotor / Fan Anomaly Detection from Vibration | Edge Impulse · anomaly block · sensor logs |
| 25 | AnomalyTemperature / Humidity Drift Anomaly on IoT Node | Edge Impulse · time-series sensors |
| 26 | AnomalyBearing Fault Detection with Spectral Features | Edge Impulse · FFT · vibration dataset |
| 27 | AnomalyCurrent / Power Anomaly for Smart Plug MCU | Edge Impulse · current sensor data |
| 28 | AnomalyAcoustic Anomaly Detection in Industrial Equipment | Edge Impulse · unsupervised audio |
| 29 | AnomalyComparative Study: Supervised vs Anomaly Blocks | Edge Impulse · labeled + unlabeled sets |
| 🔧 Deployment · Hardware · Optimization | ||
| 30 | DeployEnd-to-End Deploy: Collect → Train → Arduino Nano 33 | Edge Impulse · Arduino library |
| 31 | DeployESP32 TinyML Pipeline with Wi-Fi Result Reporting | Edge Impulse · ESP-IDF / Arduino |
| 32 | DeploySTM32 Deployment via Edge Impulse + Cube.AI Concepts | Edge Impulse · STM32 boards |
| 33 | DeployNordic nRF / Sense Board Continuous Inferencing | Edge Impulse · Nordic SDK |
| 34 | DeployRaspberry Pi Pico (RP2040) Edge Impulse Firmware | Edge Impulse · Pico SDK |
| 35 | DeployLatency / RAM / Flash Benchmark across Boards | Edge Impulse · multi-board metrics |
| 36 | DeployINT8 Quantization Impact on Accuracy and Size | Edge Impulse · EON Tuner |
| 37 | DeployEON Compiler / Optimizations for Flash Budget | Edge Impulse · EON · profiling |
| 📡 Sensor Fusion · Multi-Modal · Continuous | ||
| 38 | FusionIMU + Audio Fusion for Context-Aware Wearable | Edge Impulse · multi-sensor project |
| 39 | FusionCamera + IMU Fusion for Gesture Robustness | Edge Impulse · synchronized datasets |
| 40 | FusionEnvironmental Sensor Fusion (Temp, Hum, Gas, Sound) | Edge Impulse · multi-axis data |
| 41 | FusionContinuous Learning / Incremental Update Concepts | Edge Impulse · transfer / retrain flow |
| 42 | FusionWindowing and Overlap Strategy for Streaming Sensors | Edge Impulse · DSP pipeline |
| 43 | FusionMulti-Label Classification for Concurrent Events | Edge Impulse · multi-label datasets |
| 🏠 Applications · Smart Home · Health · Industry | ||
| 44 | AppSmart Home Keyword Control (Lights / Fan / Door) | Edge Impulse · KWS · relays |
| 45 | AppOccupancy Detection for Room Energy Saving | Edge Impulse · PIR + vision |
| 46 | AppWildlife / Bird Call Classification on Solar Node | Edge Impulse · bioacoustics set |
| 47 | AppHand Hygiene Compliance Monitoring Wearable | Edge Impulse · IMU gestures |
| 48 | AppSmart Agriculture Soil / Leaf Condition Edge Node | Edge Impulse · sensor + vision |
| 49 | AppFactory PPE Compliance Camera on MCU | Edge Impulse FOMO · PPE images |
| 50 | AppParking Slot Occupancy with Ultrasonic + Classifier | Edge Impulse · ultrasonic features |
| 51 | AppSmart Kitchen Appliance Sound Event Detection | Edge Impulse · kitchen audio set |
| 📊 Datasets · DSP · Evaluation · Best Practices | ||
| 52 | DataCustom Dataset Collection Best Practices in Edge Impulse | Edge Impulse Studio · labeling |
| 53 | DataMFCC / Spectrogram / Spectral Features Comparison | Edge Impulse DSP blocks · audio set |
| 54 | DataData Augmentation Strategies for Tiny Audio / IMU | Edge Impulse · augmentation tools |
| 55 | DataClass Imbalance Handling for Rare Event Detection | Edge Impulse · sampling strategies |
| 56 | DataOn-Device vs Studio Accuracy Gap Analysis | Edge Impulse · confusion matrices |
| 57 | DataPublic Edge Impulse Project Replication and Improvement | Edge Impulse public projects |
| 🔬 Advanced · Research · Systems | ||
| 58 | AdvTransfer Learning from Public Edge Impulse Models | Edge Impulse · pretrained blocks |
| 59 | AdvFederated / Multi-Device Dataset Aggregation Concepts | Edge Impulse · multi-device collect |
| 60 | AdvSecure Model Update and OTA Concepts for TinyML | Edge Impulse · firmware update notes |
| 61 | AdvPower Measurement of Inference vs Idle for Battery Life | Edge Impulse · current probe data |
| 62 | AdvBenchmark: Edge Impulse vs Hand-Crafted TFLite Micro Pipeline | Edge Impulse · TFLite Micro |
| 63 | AdvCapstone: Full Product — Collect → Train → Deploy → Dashboard | Edge Impulse · MCU · full stack |
Topics follow Edge Impulse Studio practice. Contact us for project setup notes, dataset guidance, deployment steps, university-format report, PPT and viva Q&A.
Why Choose Us for Edge Impulse TinyML Projects?
Bangalore guidance for students deploying on-device ML with Edge Impulse.
Audio & Keyword Spotting
Wake-word and sound event models with MFCC/spectrogram DSP blocks and continuous inferencing.
FOMO Vision
Person/object detection and counting optimized for MCU cameras and tight flash budgets.
IMU & Anomaly
Gesture, activity, fall detection and vibration anomaly blocks with real sensor logs.
Multi-Board Deploy
Arduino, ESP32, STM32 and Nordic firmware with latency, RAM and power metrics.
FAQ — Edge Impulse TinyML Projects
Strong topics include keyword spotting, FOMO person detection, IMU gesture/fall detection, vibration anomaly detection, and full collect–train–deploy pipelines on Arduino Nano 33 or ESP32 with clear memory and latency reporting.
Edge Impulse Studio, FOMO, DSP blocks, EON Tuner; boards Arduino/ESP32/STM32/Nordic; datasets Google Speech Commands, ESC-50, HAR sets and custom collected audio/IMU/image data.
Yes. Packages include Studio project guidance, dataset notes, deployment steps, university-format report, PPT and viva Q&A.
Edge Impulse TinyML Lab — Bangalore
Data collection, training and MCU deployment support for TinyML projects.
Keyword Spotting
Audio Lab
KWS · MFCCAudio Lab
FOMO Vision
on MCU Camera
Person · Counton MCU Camera
IMU Gesture
& HAR
Wearable Motion& HAR
Anomaly
Detection
Vibration · AudioDetection
Arduino · ESP32
STM32 Deploy
Multi-BoardSTM32 Deploy
EON Tuner
Optimization
Flash · LatencyOptimization
Custom Dataset
Collection
Studio LabelingCollection
Viva & Report
Support
DocumentationSupport