Professor Alessandro Cunha
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
1
Status Quo - AI/ML in the cloud
Latency and non-real-time response
Security and data privacy risks
Inefficient power – requires radio wakeup
Relatively high cost
AI/ML Processing power only available in the cloud
Inference in the cloud
Sense at the edge
Limitations of cloud AI/ML
Internet connection required
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
2
PSOC™ Edge - AI/ML at the Edge!
Low latency and real-time response
Improved security and data privacy
Optimized power efficiency
Reduced cost
Sense and Inference at the edge
Key benefits of Edge AI/ML
No internet connection required
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
3
PSOCTM Edge – Focus on Consumer & Industrial applications, �AI/ML Use Case Compute Requirements for Next Generation Devices
HMI
Smart Home
Robotics
PSOC™ Edge E8x
Wearables
Security Camera
Audio, Voice & Language
Vision
Device Management
DEEPCRAFT™ Studio + ModusToolbox™ provide the full journey from ML model development to embedded software
System solution offering with other Infineon products
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
4
Enabling Responsive Edge Devices
Human Machine Interface (HMI)
Smart Home
Robotics
Wearables
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
5
�Unlock high-end user experiences for next gen devices with PSOCTM Edge Software & ML Ecosystem
Connect
Security
Voice �& Audio
HMI
Machine Learning
Vision
Full unified set of traditional embedded Development Tools – provided through ModusToolboxTM development environment.
Comprehensive set of voice products (voice assistant, wake word, acoustic front end, ..) offering lowest power without compromising user experience
DEEPCRAFT™ Studio ML Development Tool & ModusToolboxTM ML for data collection & pre-processing, �model training, model conversion & deployment
Ready ML Models adding specific capability immediately
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
6
PSOCTM Edge expands AI/ML use cases of our PSOC 6 portfolio�AI/ML Use Case Compute Requirements for Next Gen Devices
Device Management
Vision
(Camera, RADAR)
Audio, Voice & Language
PSOC 6 Series�
PSOC Edge E81�
PSOC Edge E83/E84�
Anomaly Detection
Predictive Maintenance
Autonomous Operation
ML Based Control
Beamforming
KWS / Simple Commands
1 Mic Noise Suppression
Voice Prompts
Natural Language Processing
Acoustic Event Detection
Natural Commands
Advanced ML Voice Enhancements
Audio Playback Support
2 Mic Noise Suppression
Movement / Presence Detection
Person Detection
Face Recognition
Object Detection
Position Detection
Gesture Detection
Surface Detection
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
7
PSOC™ Edge – Next Gen MCU based Edge Device Platform
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
8
PSOCTM Edge – Ecosystem for AI/ML Applications
PSOCTM Edge Software Blocks
PSOCTM Edge SW & AI Ecosystem
PSOCTM Edge Embedded Development
Portfolio of modular, interoperable SW building blocks for evaluation and productive use
ModusToolboxTM & DEEPCRAFTTM Studio
Development Board
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
9
PSOC™ Edge E8x Evaluation Kits
Kit Name | Purpose | MPN | Availability |
PSOC™ Edge E84 Evaluation Kit | PSOC™ Edge Device functional evaluation | KIT_PSE84_EVAL | July 2025 |
PSOC™ Edge E84 SOM | PSOC™ Edge Minimal System Module | MOD_PSE4_SOM | TBD |
PSOC™ Edge E84 AI Kit | Evaluate AI capabilities of PSOC™ Edge using a low-cost kit featuring multiple sensors | KIT_PSE84_AI | July 2025 |
PSOC™ Edge E84 Smart HMI Kit | Smart HMI Solution Kit to showcase AI capabilities with multiple sensors, capacitive touch and graphics | TBD | TBD |
PSOC™ Edge E84
Evaluation Kit
PSOC™ Edge E84
SOM
PSOC™ Edge E84
AI Kit
PSOC™ Edge E84
Smart HMI Kit
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
10
PSOC™ Edge E84 AI Kit�KIT_PSE84_AI
Kit Status
Kit Content
Highlights
Featured Components
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
11
PSOC™ Edge E84 AI Kit�KIT_PSE84_AI
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
12
PSOC™ Edge E84 AI Kit�KIT_PSE84_AI
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
13
DEEPCRAFT™ Studio + ModusToolbox™ provide the full journey from ML model development to embedded software
DEEPCRAFT™ Studio
ModusToolboxTM
Bring your own data
Optimize & validate your model
Develop your embedded product with IFX MCUs
Bring your own model
Buy a Ready Model
OR
Data Acquisition
Preprocessing
Model Training
DEEPCRAFT™ Studio + ModusToolboxTM
Optimization
Deployment
.c
.a
Data �Import
Data �Capture
Data �Generation
Data �Labeling
Data �Augmentation
Preprocessor / Feature extractor selection
Model�Selection
Model�Training
Model�Validation
Hardware Optimization
Code�Generation
Target Device �Test and Validation
Field �Deployment
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
14
DEEPCRAFT™ Studio + ModusToolbox™ provide the full journey from ML model development to embedded software
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
15
DEEPCRAFT™ Studio helps you take your edge AI �ideas to production quickly and easily
State-of-the art, end-to-end ML development platform: Collect & annotate data directly from your target hardware. Create, train, evaluate & deploy great ML models fast.
Own your own data. Data is only used by Imagimob to train your models. Data is stored offline on your machine.
Not locked into the ecosystem: Build a custom model, or bring your own to optimize for the edge, and deploy on the hardware of your choice.
AutoML functionality: Auto-generates high performance AI models optimized for speed and low footprint.
Visualization is king: No more "black box": Follow your machine learning model creation journey with our Graph UX.
Collect & annotate high quality data
Manage data into different datasets
Build & train great models
Evaluate and find the best model
Optimize and package application
Deployment and Maintenance
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
16
Solar panel failure detection�Demo Owner: Ella Rickerson (CSS ICW AM MSM) / Omar Cruz (CSS ICW MCU PMG)
Demo Description
The demo helps in the increase of the efficiency during the energy production and the detection of failures in the energy production from solar panels. These failures could have a physical failure of the system as origin or a fraud from the solar panel owner.
This demo aims to implement a Maximum Power Point Tracking (MPPT) algorithm enhanced with Artificial Intelligence (AI) techniques to optimize the efficiency of solar panels. Using a PSoC Edge microcontroller and Infineon DC-DC converters, the system will adjust the converter's duty cycle in real-time to maintain operation at peak efficiency, considering variations in temperature and irradiance
Infineon Products Used
Silicon: PSoC Edge E84, CYW55513IUBG Wi-Fi & BLE, TLE4971 Current Sensor
Hardware PSoC Edge E84 SoM Eval Kit, Custom IFX DC-DC converter PhD
Software: ModusToolbox IDE
Demo setup requirements
Size: 105 x 40 x 10 mm
Power: USB-C Power Supply Interface
Internet connection: Bluetooth Low Energy (BLE)
Other equipment needed: Android Tablet
Target Applications
Residential solar energy systems, Solar panel manufacturers
Commercial and industrial solar installations
Integration in IoT devices for smart energy management.
Feature | Benefit |
AI-enhanced MPPT algorithm | Fast and precise optimization of the maximum power point, improving energy efficiency |
PSoC Edge with hardware-accelerated ML | Efficient and fast data processing, reducing the system's power consumption |
Integration of BLE | Real-time remote monitoring and control, facilitating management and maintenance |
Anomaly monitoring and fault detection | Enhances system reliability by proactively identifying and correcting faults |
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
17
Copyright © Infineon Technologies AG 2024. All rights reserved.
restricted
2024-03-19
18