Nvidia Corp Enhances Humanoid Robotics Safety for Human Interaction
Nvidia Corp is dedicating efforts to render humanoid robots more secure in proximity to humans, asserting that these machines must adeptly execute split-second decisions prior to gaining human trust.
The technology titan is providing software and semiconductors to facilitate the integration of humanoids into diverse work environments, promoting genuine interaction—including physical contact when required.
Nvidia’s Halos software, derived from algorithms designed for self-driving vehicles, will underpin computational systems, enabling robots to significantly augment their situational awareness, as the company articulated in a statement dated June 22.
Nvidia, alongside other innovators in Silicon Valley, is in a fervent race to pioneer advancements in robotics, marketing this sector as the subsequent monumental frontier for artificial intelligence. Industry leaders anticipate a burgeoning market comprising billions of devices.
The humanoid robotics segment—characterized by machines that mimic human anatomy, featuring heads, torsos, and limbs—is forecasted to generate a staggering US$200 billion (RM829.8 billion) in revenue by the year 2035, according to estimates provided by Barclays.
Nonetheless, the industry grapples with significant challenges. Current safety protocols dictate that robots must either halt or decelerate upon detecting potential contact with humans, consequently diminishing their productivity.
Additionally, these constraints impede collaborative efforts, which may include transferring items to coworkers or assisting in lifting heavy loads.
“When considering safety within the framework of traditional robotics, the paradigm typically necessitates caging or deploying sensors that halt motion upon detecting obstacles,” elucidated Amit Goel, a senior director of product management at Nvidia.
“However, this paradigm falls short in addressing the complexities associated with humanoid robots.”
Nvidia plans to furnish technology for robots such as Agility Robotics’ Digit, enabling autonomous decision-making through rigorous environmental analysis.
This evolution necessitates specialized hardware conducive to safety awareness. The firm proposes Halos as an operating system interfacing with its IGX Thor hardware.
The robots will also benefit from connections to external sensor arrays. For instance, an autonomous forklift could utilize warehouse cameras to ascertain surroundings prior to navigating around corners, effectively determining whether to maintain speed or decelerate to avert a collision, the company confirmed.
As the world’s preeminent company by market capitalization, Nvidia primarily derives revenue from chips utilized in AI data centers.
Nevertheless, it has been progressively diversifying its portfolio—often providing technology at no cost—to extend the applicability of artificial intelligence.
In a further initiative to expedite robotics assimilation, Nvidia has established a laboratory enabling robot manufacturers and clients to conduct safety assessments prior to approaching regulatory bodies for required certifications.
Engineers at Nvidia will carry out preliminary inspections and assist with necessary engineering adjustments.
The parameters governing robots diverge from those that govern autonomous vehicles, which predominantly focus on avoiding contact with humans and other objects.
Humanoid robots necessitate a higher degree of adaptability, stated Pras Velagapudi, chief technology officer at Agility Robotics.
This company’s humanoids have already been deployed at Toyota Motor Corp’s manufacturing plants in Canada.
“Safety design must be considerably more sophisticated since it involves reasoning based on environmental perception regarding what can and cannot be touched, moved, or exerted force upon,” Velagapudi commented.
“Moreover, the magnitudes of these forces cannot be negligible. One cannot sidestep the dilemma by rendering the robot excessively weak, as this would inhibit its capacity to perform meaningful tasks.”
While engineers are still grappling with the technical obstacles of creating robotic beings, the vast potential of this industry allows for incremental learning.
According to Velagapudi, machines tasked with assisting individuals in domestic settings will emerge subsequently, predicated on advancements made by factory robots.
“We are initiating our endeavors within warehouse and logistics—a highly structured environment that already accounts for billions of dollars in addressable market value,” he noted.
“Outreach into retail, healthcare, and construction sectors will soon follow. Each of these presents unique challenges, yet each represents a massive industry poised to support an extensive array of robots.”
Source link: Thestar.com.my.
