A startup is taking a different route in the mobile robotics business, using an AI foundation model and a low-cost passive sensor rather than LiDAR technology, which is the backbone for most autonomous mobile devices, including self-driving cars and buses.

Vayu Robotics, which came out of stealth in October with $12.7 million in seed money from Khosla Ventures and other investors, on Tuesday unveiled its first on-road delivery robot, which company executives said is not only less expensive and simpler to operate, but also can carry packages weighing up to 100 pounds, move at almost 20 mph, and unload the packages on driveways or porches.

Such benefits come from Vayu ditching the LiDAR technology, according to Vayu Robotics CEO Anand Gopalan.

“The unique set of technologies we have developed at Vayu have allowed us to solve problems that have plagued delivery robots over the past decade and finally create a solution that can actually be deployed at scale and enable the cheap transport of goods everywhere,” Gopalan said in a statement.

Online shopping only promises to grow, according to Vayu executives, who said that by 2027, 23% of retail purchases in the United States will take place online. The problem is that the cost-per-deliver continues to be high. The technology the company is putting into its delivery robot in place of LiDAR will help drive that down, according to Gopalan.

A Less-Expensive Technology

LiDAR – or Light Detection and Ranging – is the remote sensing technology of choice for many companies working to build mapping and other capabilities into their autonomous vehicles, including delivery robots. The advantages of the technology include high accuracy, fast acquisition and processing of data, and little human intervention. And it’s big business, with Presence Research analysts expecting the global LiDAR market, at $1.95 billion last year, to hit $13.74 billion by 2033.

However, LiDAR sensors are expensive and come with software modules that are built to do one job at a time, making them inflexible and unable to adapt to new scenarios, Gopalan said. Vayu’s technology addresses those drawbacks. The company’s transformer-based mobility foundation model, Vayu Drive, uses simulation-first approach that executives said enables representation learning, which takes raw data and transforms it into understandable patterns that the model can use for classification and predictions.

The approach also allows for improved behavior planning and a faster iteration cycle.

The Model and the Sensor

The foundation model works with Vayu Sense, the lightweight passive sensor that combines AI with plenoptic sensing – which measures image brightness and the direction of light rays – to create the robot’s vision system that can manage high-resolution depth and precision, detect small objects, recognize negative space, and work under challenging conditions like low light, glare, fog, and reflections from glass and water.

This lets Vayu’s delivery robot run autonomously without having to pre-map the roads in will drive on and also can navigate in stores and on city streets, according to Vayu.

With the growth of ecommerce and the continued innovation of generative AI and machine learning, the use of autonomous package delivery is expected to grow rapidly.

The Need for Autonomous Delivery

“The logistics and delivery industry is undergoing a technology-driven transformation, with robotics, drones, and autonomous vehicles expected to play a key role in meeting the growing challenges of last-mile delivery,” researchers wrote in a report published last year in Nature. “The last-mile logistics of freight distribution is a critical and challenging aspect of the supply chain. It is often the least efficient and most expensive stage, accounting for up to 28% of delivery transportation costs, and is a considerable source of congestion and other externalities, such as pollution and noise.”

All of these inefficiencies and the harm to people are making last-mile logistics a priority for businesses and policymakers, the researchers wrote. That’s helping to drive major retailers and logistics companies like Amazon, Walmart, and UPS to look at autonomous delivery. That can be seen on the money spent, with the global market for delivery robots expected to grow from $236.6 million in 2022 to more than $2.1 billion by 2030, according to Fortune Business Insights.

An Experienced Team

Vayu brings a lot of experience, with co-founder Gopalan being the CEO of Velodyne, a LiDAR supply that went public in 2020. Mahesh Krishnamurthi, a co-founder and chief product officer at Vay, came from Apple Special Projects Group (SPG) and Lyft, while co-founder and CTO Nitish Srivastava also came from Apple SPG and the AI lab at the University of Toronto. Geoffrey Hinton, who started the lab and now is an advisor to Vayu.

The company’s delivery robots already are being used, with an ecommerce company signing an agreement to use 2,500 to deliver goods. Similar customers are in the works, according to Vayu. At the same time, the startup is working with a global robotics manufacturer to replace LiDAR sensors with Vayu’s technology to be used for the robotic applications.

Gopalan said the plan is to expand the presence of its technology.

“Our software is robot form factor-agnostic and we have already deployed it across several wheeled form factors,” he said. “In the near future, Vayu’s software technology will enable the movement of quadrupedal and bipedal robots, allowing us to expand into those markets as well.”