Add Pi0.5 with RTC pipeline in Robotic AI Suites

[baihe-liu]

Description

This pull request adds a new sample pipeline for the Pi0.5 Vision-Language-Action (VLA) model with Real-Time Chunking (RTC) optimized for the OpenVINO toolkit, targeting robotics applications. The integration includes documentation, usage instructions, benchmarking, and evaluation scripts, as well as the addition of the lerobot submodule for model support.

Key changes:

New Pipeline Integration and Documentation

• Introduced a new sample pipeline, "Pi0.5 with Real-Time Chunking (OpenVINO Toolkit)", to the Robotics AI Suite. This includes a detailed overview of the Pi0.5 model architecture, the RTC inference strategy, and the synergy between the two for high-frequency, low-latency robotic control. The documentation covers setup, model conversion, benchmarking, and both simulation and real-robot deployment instructions.
• Added the pipeline to the main list of available pipelines in README.md, providing a brief description and links for further details.
• Included the new pipeline in the sample pipelines index for the documentation.

Any Newly Introduced Dependencies

Submodule Addition

• Added the lerobot submodule at robotics-ai-suite/pipelines/pi05-rtc-ov/lerobot, referencing the Hugging Face repository, to provide the necessary model codebase and facilitate OpenVINO integration and RTC support. The license of lerobot is apache-2.0.

How Has This Been Tested?

The modification has passed coverity scan, bandit scan and functional validation.

Checklist:

• I agree to use the APACHE-2.0 license for my code changes.
• I have not introduced any 3rd party components incompatible with APACHE-2.0.
• I have not included any company confidential information, trade secret, password or security token.
• I have performed a self-review of my code.