Multi-class motion detection for PIR based building security with >98% accuracy
Single PIR sensor based solution for differentiating between multiple sources of motion leverarging a MCU-scalable Edge AI solution.
Application overview
Traditional PIR systems often misinterpret pets, vehicles, or environmental changes as human motion leading to false triggers. TI’s Edge AI based motion detection introduces intelligence at the sensor level, enabling accurate recognition and isolation of human presence from other sources of motion.
Powered by TI’s proprietary PIR motion detection model and SimpleLink™ Wireless MCUs, the CC2755 family delivers up to 5× faster inference using its integrated NPU-CDE acceleration core — achieving reliable, low-power sensing for next-generation building security systems.
Starting evaluation
Data collection
Capture accurate PIR data with the TIDA-010997 Edge AI Sensor BoosterPack™, designed for seamless integration across multiple TI LaunchPad™ platforms.
Data quality assessment
High-quality training data is the foundation of any successful AI solution. To help ensure your collected data represents the full range of real-world conditions and to be able to tag and identify key features across different sources of motions we use a Goodness of Fit (GoF) evaluation.
This GoF methodology assesses whether your dataset sufficiently captures the necessary scenarios and variability for accurate model training. Learn more about our GoF approach and see how it ensures your edge AI model is built on data that truly matters. If you prefer to get started right away, you can leverage our existing sample dataset within the CCStudio™ Edge AI Studio development tool.
Figure 1.1 - Poor data quality
Figure 1.2 - Well-separated hyper-planes for human vs non-human
GoF test illustrating the influence of pre-processing feature extraction steps on data separability.
Build and train your model
With your dataset ready, explore, train, and evaluate models using CCStudio Edge AI Studio development tool. This intuitive, GUI-based tool simplifies the entire process of model creation and deployment. Get started quickly with the PIR Motion Detection example project — complete with a preloaded dataset.
Prefer custom data? The platform includes integrated data capture and hosting tools within TI Edge AI Studio.
Find the right model for your needs
TI provides a collection of optimized PIR motion detection models tailored for different performance and power requirements. Expolore the TI tinyML tensor lab user guide to discover models that best fit your application.
Model Flexibility to support your application needs based on the memory constraints and compute requirements.
Deploying your model
Once trained and validated, deploy your model to TI’s SimpleLink™ Wireless MCUs and experience real-time performance directly on our hardware.
Edge AI Studio software development tools streamlines deployment on a end-to-end process, while the SimpleLink SDK EdgeAI Plugin enables deeper customization and integration into your embedded applications.
Choosing the Right device for you
The SimpleLink™ Wireless MCU family delivers scalable performance for executing PIR motion detection models efficiently, combining AI acceleration, low-power operation, and advanced wireless connectivity to enable a truly intelligent sensing ecosystem.
| Product number | Processing core | NPU available | Clock frequency (MHz) | PIR benchmarking metrics | ||
|---|---|---|---|---|---|---|
| Latency (ms) | Flash (kB) | SRAM (kB) | ||||
| CC2755R10 | Arm®Cortex®- M33 | Yes | 96 | 24.5 | 75 | 23 |
| CC1352R | Arm®Cortex®- M33 | No | 48 | 134 | 89 | 31 |
| CC1352P7 | Arm® Cortex®- M4F | No | 48 | 134 | 89 | 31 |
| CC1354P10 | Arm®Cortex®- M33 | No | 48 | 62.5 | 82 | 29 |
| MSPM0G5187 | Arm® Cortex®-M0+ | Yes | 80 | 5.25 | 64.1 | 10.4 |
| MSPM33C321A | Arm®Cortex®- M33 | No | 160 | 19.82 | 201 | 15 |
All the hardware, software and resources you’ll need to get started
Hardware
LP-EMCC2745R10-Q1
Evaluation and development board for SimpleLink™ 2.4GHz wireless MCUCC27x5R10 devices, that in conjunction with TIDA-010997 enables the AI-powered PIR Motion detection application.
TIDA-010997
Reference design of a sensory booster pack that features a 2x PIR AFE for PIR motion detection.
LP-EM-CC1354P10
Evaluation and development board for CC1354P10 LaunchPad™ development kit for SimpleLink™ Sub-1 GHz and 2.4GHz wireless microcontroller, that in conjunction with TIDA-010997 enables the AI-powered PIR motion detection application.
LP-CC1352P7
Evaluation and development board for CC1352P7 LaunchPad™ development kit for SimpleLink™ multi-band wireless MCU, that in conjunction with TIDA-010997 enables the AI-powered PIR Motion detection application.
LAUNCHXL-CC1352R1
Evaluation and development board for CC1352R LaunchPad™ development kit for SimpleLink™ multi-band wireless MCU, that in conjunction with TIDA-010997 enables the AI-powered PIR Motion detection application.
LP-MSPM33C321A
An easy-to-use evaluation module for the MSPM33C321A MCU. The LaunchPad kit contains an onboard debug probe for programming, debugging, and EnergyTrace™ technology. Use in conjunction with MSPM33 software development kit and TI’s Edge AI Studio to develop AI enabled PIR motion classification models.
LP-MSPM0G5187
A user-friendly evaluation module for the MSPM0G5187 MCU. The LaunchPad kit includes an onboard debug probe for programming, debugging, and EnergyTrace™ technology. Use in conjunction with MSPM0 software development kit and TI’s Edge AI Studio to develop AI enabled PIR motion classification models.
Software & development tools
CCStudio™ Edge AI Studio
A fully integrated no-code solution for training and compiling PIR Motion detection models, to deploy onto TI embedded microcontroller devices.
SIMPLELINK SDK edge AI plug-in
The SimpleLink™ SDK edge AI plug-in is a companion software package that enables advanced artificial intelligence functionality on various SimpleLink™ MCU platforms.