Real-time risk reasoning
AssigneesToyota Motor Engineering & Manufacturing North America, Inc., Toyota Jidosha Kabushiki Kaisha
InventorsYunsheng Ma, Amr Abdelraouf, Rohit Gupta, Kyungtae Han
Life of this patent
Estimating the term…
Abstract
Systems, methods, and other embodiments described herein relate to real-time risk reasoning within a vehicle. In one embodiment, a method includes acquiring sensor data about a surrounding environment of a vehicle. The method includes, responsive to identifying a risk to the vehicle from the sensor data, merging tokens that are features extracted from the sensor data into a condensed representation of a context of the surrounding environment associated with the risk. The method includes providing a driving recommendation according to the risk.
Claims20
A risk system, comprising:
- one or more processors;
- and
- a memory communicably coupled to the one or more processors and storing:
- a control module including instructions that, when executed by the one or more processors, cause the one or more processors to:
- acquire sensor data about a surrounding environment of a vehicle;
- responsive to identifying a risk to the vehicle from the sensor data, merge tokens that are features extracted from the sensor data into a condensed representation of a context of the surrounding environment associated with the risk;
- and
- provide a driving recommendation according to the risk, including instructions to perform retrieval-augmented generation (RAG) to analyze the condensed representation according to a long-term memory implemented by the RAG, the long-term memory storing previous occurrences of risks and the instructions to perform the RAG including instructions to query the long-term memory for contextual information about the surrounding environment,
- wherein the instructions to perform the RAG generate additional tokens for the contextual information and integrate the additional tokens with the tokens extracted from the sensor data.
A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:
- acquire sensor data about a surrounding environment of a vehicle;
- responsive to identifying a risk to the vehicle from the sensor data, merge tokens that are features extracted from the sensor data into a condensed representation of a context of the surrounding environment associated with the risk;
- and
- provide a driving recommendation according to the risk, including instructions to perform retrieval-augmented generation (RAG) to analyze the condensed representation according to a long-term memory implemented by the RAG, the long-term memory storing previous occurrences of risks and the instructions to perform the RAG including instructions to query the long-term memory for contextual information about the surrounding environment,
- wherein the instructions to perform the RAG generate additional tokens for the contextual information and integrate the additional tokens with the tokens extracted from the sensor data.
A method, comprising:
- acquiring sensor data about a surrounding environment of a vehicle;
- responsive to identifying a risk to the vehicle from the sensor data, merging tokens that are features extracted from the sensor data into a condensed representation of a context of the surrounding environment associated with the risk;
- and
- providing a driving recommendation according to the risk, including performing retrieval-augmented generation (RAG) to analyze the condensed representation according to a long-term memory implemented by the RAG, the long-term memory storing previous occurrences of risks and performing the RAG including querying the long-term memory for contextual information about the surrounding environment,
- wherein performing the RAG generates additional tokens for the contextual information and integrates the additional tokens with the tokens extracted from the sensor data.
Dependencies are inferred from the claim text.