Detector, system and method for detecting vehicle lock status

A system for detecting vehicle lock status, comprising:

• a detector designed and sized to be attached to a vehicle door for detecting vehicle lock status, wherein the detector comprises:
• a motion sensor configured to sense vehicle door open and close movement;
• a microphone configured to record sound patterns generated with targeted events;
• a wireless communication module;
• a memory for storage;
• a battery for supplying power;
• and
• a controller connected to the motion sensor, microphone, wireless communication module, memory and battery respectively and configured to switch on or off the microphone and the wireless communication module according to signals provided by the motion sensor, and control the wireless communication module to transmit sound files wirelessly;
• and
• a smart phone for wirelessly receiving the sound files from the detector and displaying the vehicle lock status, wherein the smart phone is installed with a software application for processing the sound files, detecting sound patterns to identify events and determining the vehicle lock status;
• wherein the targeted events comprise remote key lock open and close, open and close of car door, side mirror and engine, and set and release movements of parking brake, thus the detector is further capable of detecting status of the side mirror, engine and parking brake.

A method for detecting vehicle lock status, comprising:

• recording sound patterns after vehicle door open or close movement is sensed;
• transmitting sound files wirelessly;
• and
• wirelessly receiving and processing the sound files, detecting sound patterns to identify events, determining the vehicle lock status and displaying the vehicle lock status,
• the method further comprising:
• collecting different sound files with event information by vehicle manufacturer, vehicle model, year of production, users and samples as a labeled sound dataset;
• pre-processing and transforming each sound file into other data format suitable for training;
• and
• training a machine learning network by feeding the dataset as input into the network and predicting output with high accuracy to the dataset's label.