Edge Computing
Processing data near its source, on a vehicle or local server, to reduce cloud latency and dependence on continuous connectivity.
Definition
Edge computing performs work close to the data source rather than sending everything to the cloud. Industrial vehicles use onboard or site-local processing for timely decisions, continuity through wireless gaps and lower video bandwidth. Cloud services handle longer-term analysis, cross-vehicle comparison and model development.
Why Not Put Everything in the Cloud?
Network round trips cannot guarantee the response needed for close-range safety functions. Metal racks and interference create coverage gaps. Continuous multi-camera upload is expensive and often unnecessary. Local decisions with event summaries and selected clips reduce these limitations.
Cloud and Edge Responsibilities
The vehicle acquires and preprocesses sensors, runs local inference and alerts, and buffers data. The cloud aggregates fleets, analyzes long-term trends, manages reports and distributes updates. Exchange events and summaries instead of indiscriminately uploading raw streams. Local functions must have defined behaviour during disconnection.
Onboard Resource Constraints
Compute, power and cooling are limited, while vibration, dust, temperature and supply variation demand robust hardware. Models may need pruning or quantization. Measure sustained frame rate, latency, temperature and power on target hardware under real duty rather than relying on theoretical specifications.
Disconnection and Data Consistency
Size local storage for the expected outage period. Reliable timestamps and sequence numbers support replay and deduplication. Prioritize important events and throttle catch-up traffic so reconnection does not impair live functions. Synchronize clocks and record clock state, especially after power loss.
Operations and Version Management
Numerous inaccessible nodes need remote updates, software and model inventories, and health monitoring. Track connectivity, storage, restarts and resource use. These capabilities determine whether the system remains maintainable as the fleet grows and ages.
Frequently Asked Questions
Why keep safety decisions local?
Variable network latency, coverage gaps and video bandwidth make a cloud-only path unsuitable for time-critical local protection.
Is data lost when connectivity fails?
A properly sized buffer can retain data for its designed outage window and replay it later. Storage limits, clock accuracy, ordering and deduplication must all be addressed.
Related Terms
From Concept to Fleet Deployment
AI-Driven provides ADAS perception, smart onboard terminals and fleet management platforms for industrial vehicles, turning these concepts into real gains in fleet safety, efficiency and maintenance.
This glossary covers general industry concepts for technical reference and knowledge sharing. It contains no proprietary information from any specific manufacturer. For equipment specifications and operating procedures, always refer to the OEM documentation.