Photo credit: www.therobotreport.com
Ottonomy Inc., a leader in autonomous delivery solutions, has unveiled its latest innovation, Contextual AI 2.0, during the CES event. This advanced system employs vision language models (VLMs) on Ambarella Inc.’s cutting-edge N1 edge computing hardware. The development signifies a noteworthy progression toward achieving more generalized intelligence in robots, enabling the Ottobots to operate with enhanced contextual awareness and intelligent decision-making capabilities.
Amit Badlani, director of generative AI and robotics at Ambarella, remarked on the significance of this integration. He stated, “The collaboration between Ottonomy’s Contextual AI 2.0 and Ambarella’s N1 Family of SoCs represents a crucial advancement in the field of autonomous robotics. We are facilitating real-time processing and action from complex environmental data through the amalgamation of edge AI technology and VLMs.”
Ambarella’s N1 series of system-on-chips can support multiple large language models (LLMs), specifically up to 34 B-Parameters, all while maintaining low power usage. Among its offerings, the new N1-655 edge GenAI SoC features capabilities for processing a dozen simultaneous 1080p30 video streams, in addition to running various multimodal VLMs and traditional convolutional neural networks (CNNs).
Stanford University students have contributed to the advancements in real-time AI deployment by utilizing Solo Server. This technology has facilitated the implementation of VLMs, which are crucial for interpreting complex environments, as noted by Ottonomy.
Advancements in Robot Perception and Decision-Making
Ottonomy claims that its Contextual AI 2.0 has the potential to transform how robots perceive their surroundings, make decisions, and exhibit behaviors. The latest technology equips delivery robots with the ability to comprehend their environments, enabling them to detect objects while also interpreting the intricate nature of real-world scenarios.
With this heightened situational awareness, the Ottobots are designed to adapt to various environments, varying operational settings, and changes in weather and lighting conditions, according to Ottonomy.
The ability for robots to utilize contextual information rather than adhere strictly to preprogrammed behaviors is viewed by the company as a key advancement towards achieving general intelligence in robotics.
“Integrating LLMs with edge hardware serves as a pivotal advancement in our pursuit of general intelligence, allowing us to incorporate behavioral modules that leverage deep contextual insights into our Contextual AI engine,” remarked Ritukar Vijay, CEO of Ottonomy, who is scheduled to speak at 2:00 p.m. PT during the event at Mandalay Bay, Las Vegas.
Widespread Applications for Advanced Robotics
Ottonomy maintains that its approach to Contextual AI and modular design forms the backbone of its operations, as evidenced by its deployment of SAE Level 4 autonomous ground robots for a variety of tasks including the delivery of vaccines, testing kits, e-commerce packages, and spare parts across both indoor and outdoor settings within extensive manufacturing sites.
The company has established a diverse clientele spanning multiple sectors, including healthcare, intralogistics, and last-mile delivery services.
Headquartered in Santa Monica, California, Ottonomy is dedicated to innovating sustainable delivery technologies. The company is in the process of expanding its presence on a global scale, signifying its commitment to developing advanced solutions in the delivery sector.
Source
www.therobotreport.com