SimpleLink™ SDK Edge AI Plugin 9.12.00.00 Release Notes
Table of Contents
Introduction
The SimpleLink™ SDK EdgeAI Plugin (EDGEAI) from Texas Instruments is a plugin for enabling EdgeAI development on TI’s Wireless MUCs leveraging the TI’s Edge AIs Studio.
Documentation
The following documentation is provided with this release.
What’s New
This is the initial release of the Plugin. The main features that have been developed are:
- Device Agent Protocol (DAP).
- Feature Extraction Library.
- Support for Neural Processing Unit - Custom Datapath Extension (CDE).
- Smart PIR Motion Detection example.
Fixed Issues
This is the initial release of the Plugin.
Known Issues and Limitations
- Compiler support is currently limited to TI Clang.
- The PIR sensor example and the pre-trained PIR-net model in TI’s Edge AIs Studio expects a specific matrix of features as input. Therefore, the supported Feature Extraction library is not configurable in terms of features and their parameters. This will be modified in the future to enable flexibility while doing Model Training.
- The PIR sensor example and the pre-trained PIR-net model in TI’s Edge AIs Studio supports sampling frequencies between 25 and 50 Hz.
- The PIR sensor example and the pre-trained PIR-net model in TI’s Edge AIs Studio supports a fix buffer of 125 bytes as input.
- If the CC2755P10 device is used with Edge AI Studio and the model gets updated in the last deployment stage, the program will not run inference on the updated model because it always loads the model from the CC2745R10_Q1 example folder. The current solution is to manually update the CC2755P10 model with the one found inside the CC2745R10_Q1 example ai_artifacts folder.
Operating System Support
- This Plugin supports only the same Operating System as the main SDK.
Dependencies
EDGEAI requires the following dependency versions (or compatible):
- SimpleLink™ Low Power F3 SDK 9.12.00.19
- TI Clang 4.0.2.LTS-0
- Python 3.10.12
- TI’s Edge AIs Studio - Cloud Base
- TI’s Edge AIs Studio - Desktop Base EDGE-AI-STUDIO_1.5.0
- TI Neural Network Compiler for MCUs MCU_2.1.0 LTS
Device Support
This SDK supports the following SimpleLink™ wireless MCUs:
| Device | Orderable Part Number (OPN) | LaunchPad | Product Page | Migrate from these examples |
|---|---|---|---|---|
| CC2755R10 | CC2755R105E0WRHAR | N/A | CC2755R10 | LP-EM-CC2745R10-Q1 |
| CC2755P10 | CC2755P105E0WRHA | N/A | CC2755P10 | N/A |
| CC2745R10-Q1 | CC2745R10E0WRHARQ1 | LP-EM-CC2745R10-Q1 | CC2745R10-Q1 | LP-EM-CC2745R10-Q1 |
| CC2745R10-Q1 | CC2745R10E1WRHARQ1 | LP-EM-CC2745R10-Q1 | CC2745R10-Q1 | LP-EM-CC2745R10-Q1 |
Versioning
All releases have 4 segment versions (M.mm.pp.bb). This includes GA and pre-releases (engineering, alpha/EA, beta, etc.). Pre-releases are denoted with a suffix (e.g. 1.00.00.00_eng).
This product’s version follows a version format, M.mm.pp.bb, where M is a single digit Major number, mm is 2 digit minor number, pp is a 2 digit patch number, and bb is an unrestricted set of digits used as an incrementing build counter.
To support multiple side-by-side installations of the product, the product version is encoded in the top level directory.
Subsequent releases of patch upgrades will be identified by the patch number. Typically, these patches only include critical bug fixes.
Technical Support
- For support related to cc27XX devices please go to TI E2E BLE Connectivity Support Community
Last updated: 2025-11-28