The camera collects photos and video recordings. Based on the initial learning session, the trained system can detect people in the frame and assess their age, emotions, behavior or dynamics of movement (e.g. speed and direction).
After entering exact patterns as input, the system can decide whether the person can stay in the current location or whether their behavior can be treated as undesirable (e.g. under the influence of alcohol or drugs).
After teaching the system (see Machine Learning section) with precise and detailed designs of a specific product (a whole variety of fields – from electronic boards, through clothes to meat and fruit), you can use the Robotic Senses located at the place of production to detect any failures or anomalies , i.e. items not complying with the desired standard.
Then, depending on the defined process, the system can notify the operator or communicate directly with the production robot to remove or mark the defective item.
Many mechanical devices (motors, fans, transmission mechanisms) produce sounds during normal operation. The characteristics of a properly functioning device can be analyzed and given as a pattern for the system using Machine Learning techniques.
Thanks to this knowledge and sound samples collected with the help of the Robot Senses during device operation, the system can detect failures in the mechanical system or assess its condition (e.g. identify worn gears).