Sensor information processing system using container orchestration technique
An information processing device includes a container engine configured to logically allocate hardware resources on an operating system, construct a cluster for operating a container using the allocated hardware resources, and operate the container on the cluster, and an orchestration tool configured to construct a node that serves as an execution environment of a pod including one or more containers on the cluster, manage an operation of the pod on the node, and perform a restoration operation on the pod when the pod is not in a predetermined operating state. Further, the pods include a reliability calculation pod configured to calculate a reliability of each sensor based on sensor information acquired from the sensor, and a sensor information processing pod configured to perform predetermined information processing based on sensor information from one or more of the sensors calculated to have a high reliability by the reliability calculation pod.
1. A sensor information processing system, comprising:
one or more sensors; and
one or more information processing devices connected to the sensor, wherein
the information processing device includes
a container engine configured to logically allocate hardware resources on an operating system, provide a cluster for operating a container using the allocated hardware resources, and operate the container on the cluster, and
an orchestration tool configured to provide a node that serves as an execution environment for a pod including one or more containers in the cluster, manage an operation of the pod on the node, and perform a restoration operation on the pod when the pod is not in a predetermined operating state, and
the pods includes
a reliability calculation pod configured to calculate a reliability of each sensor based on sensor information acquired from the sensor, and
a sensor information processing pod configured to perform predetermined information processing based on sensor information from one or more of the sensors calculated to have a high reliability by the reliability calculation pod.
2. The sensor information processing system according to claim 1 , wherein
the sensors and the pods are provided to associate with each other in a one-to-one manner.
3. The sensor information processing system according to claim 1 , wherein
in the calculation by the reliability calculation pod, the reliability of each sensor is calculated based on a trained model generated by predetermined machine learning.
4. The sensor information processing system according to claim 3 , wherein
the trained model performs machine learning based on the reliability of each sensor calculated by the reliability calculation pod.
5. The sensor information processing system according to claim 1 , wherein
the sensor is an imaging sensor configured to acquire image information, and
the predetermined information processing in the sensor information processing pod is an object recognition process or an object detection process.
6. The sensor information processing system according to claim 1 , wherein
the pods further includes a restart process pod configured to perform a restart process on the sensor whose reliability is calculated to be low by the reliability calculation pod.
7. The sensor information processing system according to claim 6 , wherein
the restart process pod further performs a process of restarting a pod corresponding to the restarted sensor.
8. A method for controlling a sensor information processing system including one or more sensors and one or more information processing devices connected to the sensor, the method comprising:
a step of providing a container engine configured to logically allocate hardware resources on an operating system, construct a cluster for operating a container using the allocated hardware resources, and operate the container on the cluster;
a step of providing an orchestration tool configured to construct a node that serves as an execution environment for a pod including one or more containers in the cluster, manage an operation of the pod on the node, and perform a restoration operation on the pod when the pod is not in a predetermined operating state;
a reliability calculation step of calculating, by a reliability calculation pod included in the pods, a reliability of each sensor based on sensor information acquired from the sensor; and
a sensor information processing step of performing, by a sensor information processing pod included in the pods, predetermined information processing based on sensor information from one or more of the sensors calculated to have a high reliability by the reliability calculation pod.