Autonomous Conveyance Control With Real-Time Hazard Learning Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional approaches to autonomous device control systems lack the ability to incorporate shared learning from other systems, leading to inefficiencies in handling rare or unforeseen situations, and do not provide real-time updates for safety-critical requirements.
Innovation Solution
A networked system that enables real-time or near real-time sharing of learning updates across autonomous vehicles, vessels, and machines, utilizing a centralized infrastructure for generating, evaluating, and disseminating updates to all controllers, allowing for immediate adaptation to rare and unforeseen events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional control systems are used with periodic software updates, then system stability is maintained, but the ability to respond to rare or unforeseen situations is delayed
Solution Approach 1:
The system transitions from static periodic updates to dynamic continuous updates. The infrastructure continuously monitors safety-critical events from multiple controllers and immediately generates and distributes updates when new hazards are detected, rather than waiting for scheduled update cycles. This dynamic approach reduces update delivery time while maintaining system reliability.
Solution Approach 2:
The system implements continuous feedback loops where controllers report safety-critical events to the infrastructure, which then generates updated programming and feeds it back to all controllers in real-time. This feedback mechanism enables the system to learn from rare or unforeseen situations and rapidly propagate safety improvements across the entire fleet, addressing both reliability and time concerns.
2Adaptability or versatility
If individual learning is implemented in each controller, then local adaptability improves, but shared learning across the network is lost
Solution Approach 1:
The system merges individual controller learning with centralized infrastructure learning. Each controller maintains its local learning capabilities for immediate adaptation, while simultaneously contributing safety-critical event data to the infrastructure. The infrastructure aggregates this information across the network and distributes shared learning updates to all controllers, combining the benefits of both local and collective intelligence.
Solution Approach 2:
The infrastructure serves multiple functions: it acts as a centralized learning repository, a validation engine for safety updates, and a distribution network for propagating improvements. This multi-functional approach enables the system to maintain individual controller adaptability while preventing loss of shared learning information through centralized aggregation and dissemination.
3Reliability
If real-time shared learning updates are implemented, then safety response to rare events improves, but system complexity increases
Solution Approach 1:
The system segments safety-critical functions from non-critical functions. The infrastructure focuses exclusively on monitoring, validating, and distributing safety-related updates, while individual controllers handle local operational control. This segmentation reduces overall system complexity by clearly defining roles and responsibilities, enabling real-time safety updates without overwhelming the entire system.
Solution Approach 2:
The infrastructure acts as an intermediary between individual controllers and the external environment. It receives safety-critical events from controllers, validates the corresponding updates against safety requirements, and distributes approved updates to the network. This intermediary role simplifies the system architecture by centralizing complex validation and coordination tasks, reducing the burden on individual controllers while enabling real-time safety responses.
Data Source
AI summary
Disclosed subject matter identifies, characterizes, and mitigates previously unforeseen safety hazards that are likely to be encountered by autonomous conveyances—finding these hazards, assessing their potential safety impact, modifying the design to mitigate them should they occur, disseminating updated design programming to all units, including those under construction or those already in the field, and including those hazard mitigations of high severity that exceed the maximum capabilities of the controller as manufactured. These hazards can include rare, infrequent and unforeseen hazards by monitoring conveyances already in the field, gathering data from autonomous conveyances, such as those using a design being updated, and data obtained from those using other autonomous designs in the field. By obtaining data from non-autonomous conveyances, as supplied by their drivers and operators, reporting real-time via a smartphone application, categories of rare, infrequent or unforeseen hazards can be integrated into modified designs.