Simulations and AI Behind Robot Vacuums


The HPC Behind the Product Series


February 9, 2021

Robot vacuum cleaners have emerged as an application of the technological advances powered by artificial intelligence (AI) and the internet of things (IoT). They can be programmed and controlled remotely to perform their cleaning task and are smart enough to do it effectively with little user interference. Before the commercial release of robot vacuums such as iRobot and Neato brands, scores of engineers and scientists worked tirelessly to design, simulate, and manufacture such technological wonders.

The Parts that Make it Work

At least three main parts of the robot vacuum must work together: hardware, software, and sensors. We view sensors as a separate component and not hardware since they interact with the environment.

When we look at the compact hardware of an iRobot, for example, it is difficult to comprehend how much testing and computer simulation work went into its development. All we see is a plastic case that seems to be hiding an intricate mechanism that enables the robot to move and clean. 

Under the case, there are at least two layers of hardware: mechanical and electrical. They are designed to work in unison. Think of the electrical arm as the brain and the mechanical arm as the muscle. The brain commands the muscle which, in turn, gives the brain feedback and allows it to make smart decisions.

The Simulations

To optimize the performance of an iRobot, or similar robot vacuum, complex simulations, and extensive tests are performed, a few of which are discussed below.

  • Blowers’ Flow Pattern Simulation

    A smooth laminar flow through the robot’s filter will decrease pressure drop, resulting in better blower performance, and less noise. The CFD simulations optimize the airflow path as the blowers move air through the robot passages.

    The goal is to get the maximum airflow at the lowest pressure drop to achieve maximum blower efficiency for the blower operating point.  When the blower operating point is “low” on the blower PQ curve, there is a reduction in energy consumption, enabling an extended operating range for the battery.  
  • Electronics Cooling Simulations and Analysis

    All electronic components operate at high efficiency when cold. One of the undesirable effects of passing an electric current through a conductor is thermal.

    The components, usually integrated circuits powered through multi-layered printed circuit boards or PCBs, heat up during use. Proper venting and cooling remove the components’ thermal load, preventing overheating and throttling.The development team performs complex CFD-heat transfer analysis to design the heat sinks properly and strategically position components for maximum heat removal. Design verification testing is usually the final step before releasing the product to market.

Thermal contours on one of the printed circuit boards optimized using CFD software

Thermal contours of a robot vacuum
circuit board optimized using CFD software

  • Mechanical and Assembly Simulations

    One of the main concerns of the developer is structural integrity and reliability. As an example, random and harmonic vibrations reduce the lifetime of the device. The product development team performs transient and steady-state simulations and testing to identify low-frequency vibrations and add damping to the assembly. Ensuring that the system does not function close to its natural frequencies, a phenomenon known as resonance, is a top priority. 

The Magic Behind the Robot Vacuum

The original iRobot used a simple approach to navigation: it moved randomly, utilizing a series of sensors to change direction every time it encountered an obstacle such as a wall or stair. If you watch it in action, you will notice that the iRobot and the Neato vacuum cleaners perform a 360 degree rotation to determine their starting position and proximity to obstacles. 

With the release of new robot vacuums, we’ve seen added intelligence features. Using at least one camera, laser mapping in the case of Neato, and combining simultaneous location and mapping, or SLAM with sensor fusion, the latest generation robot vacuum cleaners are more efficient than their earlier variants. 

A Neato map constructed using AI software and sensor fusion.

A Neato map constructed using AI software and sensor fusion.

The addition of a camera allows the robot to capture images of a room, compare the images using AI software to gradually assemble a map of the robot’s environment, and determine its location as it performs its duties. Sensor fusion augments the robot’s camera by combining data from its proximity sensors with images taken from the camera allowing the robot vacuum to function even when it is obscured. 

Robot vacuum cleaners have come a long way since the first iRobot randomly swept the floors of techies and early adopters. Self-emptying docks allow the robot vacuum to perform more tasks, and smart mopping and cleaning are now enabled on these technological marvels. Artificial intelligence will probably take these machines to self-maintenance and become part of the interconnected house in the future.

Our blog series, The HPC Behind the Product, showcases the simulations and artificial intelligence involved in the development and design of popular consumer products including the Peloton, YETI tumblers, the MIRROR, and more.

Disclaimer: We are not affiliated, associated with, or in any way represent iRobot or Neato brands.

Subscribe to our quarterly email newsletter to get resources like this in your inbox!