Configuring and managing fleets of dynamic mechanical systems
Provided is a distributed robot management system, including: a first fleet of robots at a first facility; and a robot management server system remote from the first facility and communicatively coupled with the first fleet of robots via a network, wherein the robot management server system is configured to: provide configuration information to the first fleet of robots, maintain a remote representation of state of robots in the first fleet of robots, receive and store data from the first fleet of robots, and provide computing resources by which robots in the first fleet of robots are trained.
1 . A distributed robot management system, comprising:
a first fleet of robots; and
a robot management server system communicatively coupled with the first fleet of robots via a network, wherein the robot management server system provides multiple classes of interfaces comprising:
an administration interface for access by a system administrator, the administration interface configured to perform system administrative functions;
a partner interface for access by a first user for a first facility at which the first fleet of robots are deployable, the partner interface configured to manage tasks of the first fleet of robots at the first facility; and
a robot interface for access by individual robots in the first fleet of robots, the robot interface configured to:
request tasks;
provide a first configuration information to the first fleet of robots, wherein the first configuration information comprises hardware configuration, system configuration, and asset configuration for one or more robots in the first fleet of robots, wherein the asset configuration comprises neural network weights of a model by which the one or more robots are to be controlled,
maintain a remote representation of states of robots in the first fleet of robots, wherein maintaining the remote representation of states of robots in the first fleet of robots comprises:
providing, for at least some robots in the first fleet, manufacturing dates of the respective robots, model numbers of the respective robots, geolocations of the respective robots, current power state of the respective robots, current configuration information of the respective robots, and data indicative of tasks the respective robots are capable of performing;
receive and store data from the first fleet of robots, and
provide computing resources by which robots in the first fleet of robots are trained;
wherein a given robot in the first fleet of robots comprises a robot computer configured to operate as a client to the robot management server system via the robot interface and:
control the given robot during runtime performance of tasks;
configure the given robot in cooperation with the robot management server system via the robot interface;
train the given robot; and
provide state and experience data to the robot management server system via the robot interface;
and wherein the robot computer is configured to control more than 20 servo motors of the given robot via a reinforcement learning model configured to take as input a latent embedding space representation of a time slide of sensor data from at least five sensors of the given robot;
and further comprising a second fleet of robots, wherein the partner interface is further configured to be accessed by a second user for a second facility at which the second fleet of robots are deployable, the partner interface configured to manage tasks of the second fleet of robots at the second facility.
2 . The system of claim 1 , wherein providing the first configuration information to the first fleet of robots comprises providing respective hardware and system configurations specific to each robot model in the first fleet and each instance of the respective robot model in the first fleet.
3 . The system of claim 1 , wherein the first configuration information provided to the first fleet of robots further comprises task information and dependencies required to execute a task with robots in the first fleet.
4 . The system of claim 1 , wherein providing the first configuration information to the first fleet of robots comprises:
receiving a request for configuration information from a given robot in the first fleet of robots;
authenticating the request with public key cryptography as being from the given robot in the first fleet of robots;
determining that the given robot is authorized to access the configuration information and, in response, providing the configuration information.
5 . The system of claim 1 , wherein receiving and storing data from the first fleet of robots comprises:
receiving and storing information gained by at least one of the robots in the first fleet of robots while operating.
6 . The system of claim 5 , wherein the information gained by at least some of the robots comprises data based on sensor traces from sensors of the at least some of the robots and data indicative of a context in which the sensor tracers were obtained.
7 . The system of claim 1 , wherein a given robot in the first fleet of robots is configured to provide at least some of the data from the fleet of robots to the robot management server system via the robot interface, wherein providing the at least some of the data to the robot management server system comprises:
applying an experience filter to data indicative of runtime experiences of the given robot;
grouping the data indicative of runtime experiences into groups;
compressing the data indicative of runtime experiences; and
purging the data indicative of runtime experiences from memory of the given robot after sending the groups to the robot management server system via the robot interface.
8 . The system of claim 1 , wherein:
the robot computer is configured to host a web application or application program interface accessible to a native mobile application on a user's mobile computing device; and
the web application or native application provides access to the robot interface by which the given robot is trained or configured by the user.
9 . The system of claim 1 , wherein the given robot is coupled to an auxiliary computing device, different from the robot computer, configured to present visualizations of robot performance via the robot interface while the given robot is operating.
10 . The distributed robot management system of claim 1 wherein the robot interface is further configured to receive and deliver updates to machine learning models deployable by individual robots in the first fleet of robots.
11 . The distributed robot management system of claim 1 wherein each robot in the first fleet of robots includes a respective on-board machine learning subsystem having an encoder operative to transform high-dimensional sensor outputs into lower-dimensional vector representations in a latent embedding space.
12 . The distributed robot management system of claim 1 wherein the administration interface is further configured to enable the system administrator to control access of the first user to the first fleet of robots via the partner interface and access of the second user to the second fleet of robots via the partner interface.