Showing papers by "Warren B. Jackson published in 2020"

Current control systems and methods

Abstract: A system that includes an energy device having an active region configured to generate or consume electrical energy provided by an electrical current is discussed. A current limiter is disposed between the energy device and a current collector layer. The current limiter controls the current flow between the energy device and the current collector layer. A plurality of electrochemical transistors (ECTs) are arranged in an array such that each ECT in the array provides localized current control for the energy device. Each ECT includes a gate electrode, a drain electrode, a source electrode, and a channel disposed between the drain and the source electrodes. An electrolyte electrically couples the gate electrode to the channel such that an electrical signal at the gate electrode controls electrical conductivity of the channel. The current collector layer is a shared drain or source electrode for the ECTs.

Tamper detection for breakable remakeable connections

Abstract: A tamper detection system includes a detector that measures a value of a parameter for each connection of multiple breakable remakeable connections between first and second components. The system includes an analyzer that compares the measured parameter value for each connection or a representative value derived from measured parameter values of the connections to an expected value. Based on the comparisons, the analyzer determines whether the multiple breakable remakeable connections between the first and second components have been broken and remade.

Control of particle layer depth and thickness during powder printing

Abstract: A printing system includes a liquid ejector configured to deposit a curable layer on a surface of a substrate, the layer having a free surface and an interface between the layer and the substrate. A pre-curing device pre-cures the layer such that a first region closer to the free surface is less cured than a second region closer to the interface. The curing device includes a pre-curing initiator source configured to provide a pre-curing initiator that polymerizes the layer. The curing device also includes a pre-curing inhibitor source configured to deliver an inhibitor that inhibits polymerization of the layer. A particle delivery device delivers particles to the layer after the layer is pre-cured.

System for thread printing using image-based feedback

Abstract: A system and method for treatment of thread includes a thread printer that applies colored coatings to thread. A camera is positioned to capture images of the treated thread after it passes through an outlet of the thread printer. The system processes the images to determine whether the location of a selected segment of the treated thread corresponds to an expected location. If the actual location of the selected segment corresponds to the expected location, the system will continue instructing the thread printer to selectively apply the colored coatings to the thread according to the pattern. However, if the actual location does not correspond to the expected location, the system will modify the pattern so that thread printer adjusts the size of one or more segment to automatically account for the misalignment.

Continuous high-speed 3d printing

Abstract: A method for providing high-speed three dimensional (3D) printing is provided. The method includes producing at least one three dimensional (3D) printed part. Producing the 3D part includes continuously constructing to extend outwardly a diameter of a rotating cylindrical core via continuous deposition of a layer, and defining a first pattern in the continuously deposited layer corresponding to a cross-section of the at least one 3D printed part.

Error Autocorrelation Objective Function for Improved System Modeling

Abstract: Deep learning models are trained to minimize the error between the model's output and the actual values. The typical cost function, the Mean Squared Error (MSE), arises from maximizing the log-likelihood of additive independent, identically distributed Gaussian noise. However, minimizing MSE fails to minimize the residuals' cross-correlations, leading to over-fitting and poor extrapolation of the model outside the training set (generalization). In this paper, we introduce a "whitening" cost function, the Ljung-Box statistic, which not only minimizes the error but also minimizes the correlations between errors, ensuring that the fits enforce compatibility with an independent and identically distributed (i.i.d) gaussian noise model. The results show significant improvement in generalization for recurrent neural networks (RNNs) (1d) and image autoencoders (2d). Specifically, we look at both temporal correlations for system-id in simulated and actual mechanical systems. We also look at spatial correlation in vision autoencoders to demonstrate that the whitening objective functions lead to much better extrapolation--a property very desirable for reliable control systems.

A novel residual whitening based training to avoid overfitting.

Abstract: In this paper we demonstrate that training models to minimize the autocorrelation of the residuals as an additional penalty prevents overfitting of the machine learning models. We use different problem extrapolative testing sets, and invoking decorrelation objective functions, we create models that can predict more complex systems. The models are interpretable, extrapolative, data-efficient, and capture predictable but complex non-stochastic behavior such as unmodeled degrees of freedom and systemic measurement noise. We apply this improved modeling paradigm to several simulated systems and an actual physical system in the context of system identification. Several ways of composing domain models with neural models are examined for time series, boosting, bagging, and auto-encoding on various systems of varying complexity and non-linearity. Although this work is preliminary, we show that the ability to combine models is a very promising direction for neural modeling.

Expanded color gamut for thermochromic color processing

Abstract: An article includes a substrate with at least first and second materials disposed in or over the substrate. The first material is a non-color changeable material of at least one first color and the second material is a thermochromic color changeable material activated to produce at least one second color different from the first color. Additive color mixing of the first and second colors produces at least one third color different from the first and second colors.