A method of controlling air flow in an environmental control system includes receiving a target total flow value and a combined flow measurement at an outer control loop module. A first air flow reference value is determined based on the target total flow value and the combined flow measurement using the outer control loop module. The first air flow reference value is communicated to an inner control loop module and a flow control valve command determined based at least in part on the first air flow reference value using the inner control loop module. The flow control valve command is communicated to a flow control valve operatively associated with the inner control loop module to control flow through the ECS using the flow control valve. Environmental control systems and computer program products are also described.

A method of controlling air flow in an environmental control system includes receiving a target total flow value and a combined flow measurement at an outer control loop module. A first air flow reference value is determined based on the target total flow value and the combined flow measurement using the outer control loop module. The first air flow reference value is communicated to an inner control loop module and a flow control valve command determined based at least in part on the first air flow reference value using the inner control loop module. The flow control valve command is communicated to a flow control valve operatively associated with the inner control loop module to control flow through the ECS using the flow control valve. Environmental control systems and computer program products are also described.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing load reduction optimization. In one aspect, a method includes accessing load reduction parameters for a load reduction event, accessing energy consumption models for multiple systems involved in the load reduction event, and performing, based on the load reduction parameters and the energy consumption models, a plurality of simulations of load reduction events that simulate variations in control parameters used to control the multiple systems. The method also includes optimizing, against a load reduction curve, the load reduction event by iteratively modifying the control parameters used in the plurality of simulations of load reduction events, and outputting the optimal load reduction event with optimized control parameters.

The present invention provides systems and methods for providing cross-device native functionality for a native app. More specifically, the invention is directed to a JavaScript Object Notation (JSON) data exchange format for use with a native app running on a user's mobile device, wherein the exchange format is configured to improve user experience and interaction with the app. The present invention may be particularly useful in a mobile-based crowdsourcing platform in which data is continually exchanged between remote user devices and a cloud-based service for collecting and managing user-driven data based on user interaction with native apps on their devices.

Some embodiments relate to a method of adjusting therapeutic stimulation from a therapeutic stimulation system. The method comprising: providing therapeutic stimulation based on a plurality of stimulation parameters to a patient with an implanted controller and at least two electrodes; receiving data indicative of feedback associated with a patient rating of the therapeutic stimulation via a patient input device; automatically adjusting, with the implanted controller, at least one of the plurality of stimulation parameters; and providing adjusted therapeutic stimulation based on the adjusted at least one stimulation parameter to the patient with the controller and the at least two electrodes. The method further comprises receiving additional data indicative of feedback associated with another patient rating of the adjusted therapeutic stimulation; and executing a machine learning algorithm based on the stimulation parameters and received data indicative of feedback, to determine preferred stimulation parameters.

A prediction control device controls an actuator for automatic driving of a vehicle including: a command value generation unit generating an operation amount for the actuator and an operation amount candidate as a predicted value; an output prediction unit outputting a control amount candidate as a predicted value corresponding to the actuator output by using an operation model; an evaluation function calculation unit expressing constraint conditions for the automatic driving; a situation degree detection obtaining a measure of giving priority to ride comfort or giving priority to danger avoidance of an own vehicle while traveling; and a responsiveness adjusting unit obtaining a next operation amount candidate from the situation degree from the situation degree detection unit. The operation command value generation unit generates an operation amount for the actuator, and the responsiveness adjusting unit adjusts the output of the evaluation function according to the situation degree.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for disaggregating latent causes for computer system optimization. In one aspect, a method includes accessing a data stream for data values resulting from operations performed by a computer system; providing the data values as input to a data disaggregation machine learning model that generates descriptors of latent causes of the data values; providing the data values and the descriptors of the latent causes of the data values as inputs to a control system model that generates embedded representations of commands to modify the operations performed by the computer system; determining commands to modify the operations performed by the computer system based on the embedded representations of commands to modify the operations performed by the computer system; and providing the commands to the computer system.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for disaggregating latent causes for computer system optimization. In one aspect, a method includes accessing a data stream for data values resulting from operations performed by a computer system; providing the data values as input to a data disaggregation machine learning model that generates descriptors of latent causes of the data values; providing the data values and the descriptors of the latent causes of the data values as inputs to a control system model that generates embedded representations of commands to modify the operations performed by the computer system; determining commands to modify the operations performed by the computer system based on the embedded representations of commands to modify the operations performed by the computer system; and providing the commands to the computer system.

A novel method is proposed to operate a stylus product. In this method, inertial measurement unit (IMU) signals are used to estimate a tilted angle of the stylus product. On the other hand, acceleration signals measured when a finger taps a stylus are collected to train a deep neural network as a tap classifier. A combination of the tilted angle and the tap classifier allows a user to interact with a peripheral device (e.g. a touchscreen) by rotating and taping the stylus product. A tap classifying system and a stylus product are also provided.

A novel method is proposed to operate a stylus product. In this method, inertial measurement unit (IMU) signals are used to estimate a tilted angle of the stylus product. On the other hand, acceleration signals measured when a finger taps a stylus are collected to train a deep neural network as a tap classifier. A combination of the tilted angle and the tap classifier allows a user to interact with a peripheral device (e.g. a touchscreen) by rotating and taping the stylus product. A tap classifying system and a stylus product are also provided.