Portuguese 
English 
Spanish 
3 min of reading 
0 comments 
PINMAP Software Allows Cheap LiDAR Sensors to Detect Glass Walls with High Precision, Reducing Costs in Autonomous Robotics.
An important breakthrough promises to reduce costs and increase the efficiency of autonomous robots. Researchers have developed software capable of detecting transparent obstacles, such as glass walls, using simple and inexpensive sensors.
The technology offers performance similar to that of sophisticated equipment but at a much lower cost.
The Problem with Traditional Sensors
Autonomous robots typically use LiDAR sensors to map their environment and avoid obstacles. These sensors function like “laser eyes,” emitting light and measuring the time it takes to return. This way, they can calculate distances and identify solid objects.
-
Man uses AI to apply for 1,000 jobs while he sleeps
-
The Earth has become an orbital junkyard: 15,550 tons of space debris surround the planet with dead satellites, abandoned rockets, and fragments traveling at 28,000 km/h.
-
Unmanned and with the autonomy to cross oceans for up to 30 days, the DriX O-16 is a 15.75-meter naval drone that sails alone for 3,500 nautical miles carrying sensors for warfare, surveillance, and submarine mapping missions.
-
Solar garden table created by a Swiss company promises to generate energy at home, power everyday devices, and even pay for itself over time using only sunlight.
However, the more affordable LiDAR sensors struggle with transparent materials. Glass walls, for instance, are not detected correctly, as the laser either passes through or reflects irregularly, confusing the system. The robot may interpret the obstacle as free space and collide.
High-resolution LiDAR sensors or ultrasonic cameras can identify these obstacles. However, they are expensive and complex.
Installing this equipment can raise system costs from hundreds of thousands to millions of wons, making large-scale adoption impractical.
The Solution Proposed by the Researchers
To overcome this limitation, a team from the Daegu Gyeongbuk Institute of Science and Technology (DGIST), led by Professor Kyungjoon Park, developed the PINMAP algorithm (Probabilistic Incremental Mapping Based on Navigation). The research was published in the IEEE Transactions on Instrumentation and Measurement.
PINMAP functions innovatively. Instead of requiring new sensors, it uses the same inexpensive LiDAR sensors but processes the data differently.
These sensors, despite their limitations, can capture a few data points when interacting with transparent surfaces.
The algorithm collects these sporadic points and probabilistically calculates the presence of glass walls over time. In doing so, the system creates a reliable map of obstacles, even with incomplete and scattered data.
Technology Based on Open-Source Tools
PINMAP was developed using well-known tools in the robotics field. It utilizes Cartographer for mapping and Nav2 for navigation. Both are part of the ROS 2 ecosystem, a widely used open-source software suite in the industry.
This choice makes the solution easy to implement. There is no need to modify existing hardware or system architecture. Just apply the software to the current system, leveraging the already installed sensors.
Practical Results and Economic Gains
Tests conducted at DGIST showed impressive results. PINMAP managed to detect glass walls with 96.77% accuracy. In comparison, the traditional Cartographer-SLAM method, using the same sensors, virtually failed to identify these obstacles.
Professor Park highlighted the importance of the work: “PINMAP reverses the conventional idea that hardware performance equals system performance and proposes a new standard by which software can enhance sensor capabilities.”
In addition to technical performance, the economic benefit is significant. With PINMAP, it is possible to achieve results similar to those of high-cost sensors for less than one-tenth of the price.
This paves the way for large-scale application of service robots in environments such as hospitals, airports, malls, and warehouses.
The new technology has the potential to transform the autonomous robotics sector. By reducing the risk of collisions with transparent surfaces and keeping costs low, PINMAP can accelerate the adoption of these robots in various indoor settings, making them safer and more accessible.
The article was published in the journal IEEE Transactions on Instrumentation and Measurement.
Be the first to react!