Browse Prior Art Database

Apparatus for Machine Learning Snow Blower Assistant Disclosure Number: IPCOM000249374D
Publication Date: 2017-Feb-22
Document File: 2 page(s) / 67K

Publishing Venue

The Prior Art Database


Disclosed is a machine learning apparatus, developed for a snow blower, to facilitate efficient snow removal.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.


Apparatus for Machine Learning Snow Blower Assistant

When moving snow with a snow blower, repeatedly adjusting the chute to redirect the removed snow is inefficient and can be frustrating for the user. This becomes especially difficult after several snowstorms, as the volume of snow increases and the area to which to move the snow decreases. Another problem is that the wind often blows the snow back at the operator.

The novel contribution is a machine learning apparatus, developed for a snow blower, to facilitate efficient snow removal. The apparatus records the snow blower operator's use of the machine (e.g., position, direction, and chute position), wind amount and direction, and the amount of snow in various areas near the site where the snow is being cleared. It also reviews weather forecasts for predicted new snow accumulation and its effect on the current volume of standing snow. Standing snow may melt and evaporate, refreeze, or accumulate over multiple snow storms. Over time, the machine learns the best ways to throw the snow for minimizing clean up time, minimizing snow being blown back at the operator, and maximizing the best places to put the snow as it is moved.

In the most basic implementation, to learn the places to throw the snow, the apparatus simply logs the position of the machine, the direction of the machine, and the operator's positioning of the chute. After the system completes this learning, the user only needs to push the blower and the system repeats the learning to throw the snow in the optimal manner and direction. The blower may also be automated in its movements (with sensors to avoid running into people or new objects) so that once trained, the machine operates by repeating the learned movements.

In an assistance-oriented implementation, the apparatus learns more from sensors on the machine. The system gathers information about chute positioning, throwing direction, and weather (e.g., temperature, humidity, wind, shade, sunlight, etc.). Photographs provide information about snow accumulation/depth at various locations. Satellite photographs of the area to be cleared provide a high-level view. Sensors in the surrounding area (e.g., placed in a vertical stick at the location) track precipitation, wind direction, wind speed, temperature, humidity, sun exposure, etc. This data can provide information about snow depth and packing ability. In addition, the apparatus utilizes known art to obtain direction (e.g., compass) and positioning (e.g., Global Positioning System (GPS), iBeacons, etc.) for the machine.

The system also combines historical weather data (e.g., previous snow accumulation) and predicted weather conditions. This can help the system determine the amount of freezing and/or melting of the previously cleared and piled snow. Weather predictions help determine wind direction, wind magnitude, temperature, humidity, and precipitation for...