5BTC or 111% + 111 Free Spins!
New players only. Exclusive 111% Welcome Bonus + 111 Free Spins
Niantic technology originally built for Pokémon Go is finding a new use beyond gaming. Data gathered from millions of players is now helping delivery robots navigate city streets more precisely.
Niantic Spatial recently partnered with Coco Robotics, a company developing short distance delivery robots for food and grocery services.
Good to Know
Niantic developed a navigation system called Visual Positioning System, or VPS. Instead of relying only on GPS, the technology determines location by analyzing nearby buildings and landmarks.
Training data came largely from Pokémon Go players. Over the years, users scanned statues, streets, and public landmarks while completing in game tasks. Those scans helped build detailed 3D maps of real world locations.
Niantic says more than 30 billion images helped train the system.
“It turns out that getting Pikachu to realistically run around and getting Coco’s robot to safely and accurately move through the world is actually the same problem,” Niantic Spatial CEO John Hanke said in a recent interview with MIT Technology Review.
At its peak in 2016, Pokémon Go had roughly 230 million monthly players. Even years later, the game still maintains tens of millions of active users who continue to generate mapping data.
Delivery robots often rely on GPS for navigation. However, accuracy can drop in dense urban areas where tall buildings interfere with satellite signals.
Niantic says VPS can pinpoint location within a few centimeters by comparing camera images with its landmark database. Coco Robotics robots will use four cameras to analyze surroundings and match them with the VPS system.
“The promise of last-mile robotics is immense, but the reality of navigating chaotic city streets is one of the hardest engineering challenges,” Hanke said in a statement.
Using landmark recognition instead of GPS alone could help robots cross streets, avoid obstacles, and arrive faster with deliveries.
Pokémon Go users originally captured images to complete in game objectives. Some tasks even rewarded players for scanning monuments or public spaces through their phone cameras.
Those scans quietly built a massive visual dataset of real world locations. Niantic now uses the same data to power navigation systems outside gaming.
Crowdsourced datasets have influenced other technologies as well. CAPTCHA image tests and mapping platforms have long provided training data for machine learning systems and mapping tools.
Niantic says its long term goal centers on building a constantly updating digital map of the world. Delivery robots using VPS may add even more visual data to improve the system over time.
VPS is a navigation technology that identifies location by analyzing landmarks and surroundings rather than relying only on GPS signals.
Players scanned buildings, landmarks, and public spaces during gameplay, generating billions of images used to train the mapping model.
Coco Robotics develops small autonomous robots used for short distance delivery of food and groceries.
The system can provide more precise location data in areas where GPS signals become unreliable.
Niantic says the model was trained using more than 30 billion images collected from players.