Skyline Robotics (a company that, with the help of robotics, software engineering and mechatronics, brings together a group of creative programmers who have a shared vision of a day when robots help people with everyday tasks and employment) created “Ozmo”, a technology that combines an industrial robot, a computer vision system, artificial intelligence (AI) and machine learning technologies, to automate the operation of window cleaning.
Ozmo automatically performs tasks that previously required three to four months of manual labour. The technique also protects workers by removing them from the area or platform used for window washing.
Manual window cleaning requires two people on the platform, one on the roof and perhaps one on the ground. While the robot cleans, Ozmo only needs one person on the roof to operate the crane, electricity and water sources.
Ozmo is designed to integrate with the power and water supply infrastructure of a building maintenance unit (BMU). A table within the platform houses various types of sensors and processors.
How does it work?
The robotic arm is placed on top of the table along with its lidar camera, which uses lasers to obtain images. Once the platform is attached to the table and the building’s window washer is activated, the system is ready for operation.
Ross Blum, director of operations at Skyline Robotics says The robot arm connects to a torque sensor on the end of the cleaning brush that gives Ozmo the sensation of “touch”, as well as the Lidar camera that gives the robot “vision”.
To adapt the robot’s movements and find the most efficient cleaning path, while automating Ozmo’s glide down the side of a building, the system’s software, which acts as a brain, collects data from multiple different cameras and sensors at a rate of about 200 times/second.
To implement the window washing application for the Ozmo system, Skyline uses a software-based AI along with a waterproof version of a Kuka Robotics KR Agilus industrial robot.
The amount of pressure the robot applies to the glass is controlled by an algorithm created by Skyline that allows for feedback on force control and other factors.
Skyline chose KUKA because of the fact that KUKA’s robotic arm can withstand the rigours of an outdoor environment, and the 1.1 metre reach. Not to mention its reputation and the elite team behind it. says Blum.
The robots, which are IP65 certified, are highly reliable and vital to Skyline because of their six-axis mobility, especially when approaching the complex façade geometries found on a building.
Meeting the weight restrictions of the baskets, which differ around the world, required a major contribution from Kuka KR C5. Compared to alternative solutions, the controller is up to 35 pounds lighter and takes up much less space. At times, even with lower weight restrictions, this will allow Skyline to pack more than one arm into a single basket.
Ozmo’s world-class software and hardware components are among its many cutting-edge technological components. They allow Skyline to instantly regulate and modify the robot’s movements, as well as stabilise the system’s basket or the platform on which the robot is mounted. This is difficult because unforeseen circumstances such as wind gusts or crane problems can cause one side of the basket to drop more than the other.
Ozmo applies an opposing force to an unstable basket using the KUKA robot’s own arm to stabilise it. For a further half a second, the robot arm remains pressed against the glass of the building to prevent the basket from shaking. To ensure proper cleaning and the fastest descent path, Ozmo can process all these factors and make adjustments instantly.
Blum the Ozmo project has been under development at Skyline Robotics for the past five years, taking incremental steps to incorporate all the necessary technological components. These include the industrial robot itself and the software that allows its arm to mimic human movements to wash windows.
Skyline Robotics President and CEO Michael Brown commented; Automated window cleaning is a market that robotics companies want to enter. However, the reality is that doing so is extremely difficult and complicated, especially in terms of AI and machine learning. Partnerships with companies like Kuka, according to Skyline Robotics’ president and CEO, were crucial to the creation of Ozmo.
If you require more information, please don’t hesitate to contact us: we will be ready to help you.