Telematics Sliding Window Event Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Processing telematics data streams at scale is challenging due to duplicate event detection, where the same event is erroneously detected multiple times, especially when using sliding windows that overlap, making it difficult to distinguish correct events from duplicates.
Innovation Solution
Defining a system that uses non-overlapping logical windows within sliding windows, with each logical window being adjacent to neighboring windows but not overlapping, and including buffers to ensure events are uniquely associated with a single sliding window, preventing duplicate detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sliding windows with overlap are used for event detection, then event detection coverage is improved, but duplicate event detection occurs
Solution Approach 1:
The patent segments the sliding window into three distinct parts: a first buffer, a logical window, and a second buffer. This segmentation allows the logical window to be non-overlapping with neighboring logical windows, ensuring each event is detected only once, while the buffers provide the necessary overlap context for comprehensive event detection coverage.
Solution Approach 2:
Different parts of the sliding window structure serve different functions: the buffers (first and second) handle the overlapping regions for comprehensive coverage, while the logical window provides non-overlapping regions for unique event detection. This local differentiation resolves the contradiction between coverage and uniqueness.
2Adaptability or versatility
If sliding windows are used to process telematics data streams, then event detection capability is improved, but processing complexity increases
Solution Approach 1:
By dividing the sliding window into distinct segments (buffers and logical window), the system simplifies the management of overlapping regions. Each segment has a clear purpose, making the overall window management less complex while maintaining versatile event detection capability.
3Reliability
If overlapping sliding windows are used, then detection coverage is improved, but duplicate events are erroneously detected
Solution Approach 1:
The segmentation of the sliding window into buffers and a logical window ensures that while buffers overlap to provide comprehensive coverage, the logical windows remain non-overlapping. This prevents duplicate event detection while maintaining accurate and complete event detection coverage.
Solution Approach 2:
The patent applies different qualities to different parts of the sliding window: overlapping buffers for coverage and non-overlapping logical windows for accuracy. This local quality differentiation resolves the contradiction between detection coverage and event accuracy.
Data Source
AI summary
Systems and methods for processing telematics data streams for event detection are provided. The methods involve operating at least one processor to: receive a telematics data stream; define a plurality of sliding windows for the telematics data stream, each sliding window containing a portion of the telematics data stream, each sliding window overlapping with one or more neighboring sliding windows, each sliding window including: a logical window; a first buffer preceding the logical window; and a second buffer following the logical window; the logical window being adjacent with one or more neighboring logical windows of the one or more neighboring sliding windows without overlapping the one or more neighboring logical windows; identify a vehicle event based on the portion of the telematics data stream contained within one of the sliding windows; and store an indication of the vehicle event.