In a dual-inertia system based on permanent magnet synchronous motor (PMSM), the elastic transmission mechanism, such as reducer and coupling, is prone to mechanical resonance, which reduces the precision of servo system transmission. Model prediction and introduction of state feedback are effective methods to suppress mechanical resonance, but the control parameters de-pend on the mechanical parameters of the system. Therefore, mechanical parameter identification is very necessary. In this paper, a forgetting factor recursive least square method is used to identify the mechanical parameters of the dual-inertia elastic system. Firstly, the mathematical model of the dual-inertia system is established and reasonably simplified. Then, the discrete function of the identification algorithm is constructed, and the motor speed and electromagnetic torque are sam-pled to achieve the on-line identification of the main mechanical parameters of the dual-inertia system, and the influencing factors of the identification accuracy are analyzed. Finally, simulations and experiments are carried out to verify the effectiveness of the algorithm.

Continual learning (CL), a.k.a lifelong learning, is an emerging research topic that has been attracting increasing interest in the field of machine learning. With human activity recognition (HAR) playing a key role in enabling numerous real-world applications, an essential step towards the long-term deployment of such systems is to extend the activity model to dynamically adapt to changes in people’s everyday behavior. Current research in CL applied to HAR domain is still under-explored with researchers exploring existing methods developed for computer vision in HAR. Moreover, analysis has so far focused on task-incremental or class-incremental learning paradigms where task boundaries are known. This impedes the applicability of such methods for real-world systems. To push this field forward, we build on recent advances in the area of continual learning and design a lifelong adaptive learning framework using Prototypical Networks, LAPNet-HAR, that processes sensor-based data streams in a task-free data-incremental fashion and mitigates catastrophic forgetting using experience replay and continual prototype adaptation. Online learning is further facilitated using contrastive loss to enforce inter-class separation. LAPNet-HAR is evaluated on 5 publicly available activity datasets in terms of its ability to acquire new information while preserving previous knowledge. Our extensive empirical results demonstrate the effectiveness of LAPNet-HAR in task-free CL and uncover useful insights for future challenges.

The American Civil War has been commemorated with a great variety of monuments, memorials, and markers. These monuments were erected for a variety of reasons, beginning with memorialization of the fallen and later to honor aging veterans, commemoration of significant anniversaries associated with the conflict, memorialization sites of conflict, and celebration of the actions of military leaders. Sources reveal that during both the Jim Crow and Civil Rights eras, many were erected as part of an organized propaganda campaign to terrorize African American communities and distort the past by promoting a ‘Lost Cause’ narrative. Through subsequent decades, to this day, complex and emotional narratives have surrounded interpretive legacies of the Civil War. Instruments of commemoration, through both physical and digital intervention approaches, can be provocative and instructive, as the country deals with a slavery legacy and the commemorated objects and spaces surrounding Confederate inheritances. Today, all of these potential factors and outcomes, with internationally relevance, are surrounded by swirls of social and political contention and controversy, including the remembering/forgetting dichotomies of cultural heritage. The modern dilemma turns on the question: In today’s new era of social justice, are these monuments primarily symbols of oppression, or can we see them, in select cases, alternatively as sites of conscience and reflection encompassing more inclusive conversations about commemoration? What we save or destroy and assign as the ultimate public value of these monuments rests with how we answer this question. In this essay, drawing from the testimony of scholars and artists, I address the conceptual landscape of approaches to the presentation and evolving participatory narratives of Confederate monuments that range from absolute expungement and removal to more restrained ideas such as in situ re-contextualization, removal to museums, and preservation-in-place.

Continual learning, a crucial facet of machine learning, involves the perpetual acquisition of valuable insights from incoming data, sans the necessity for full dataset access. Esteemed as a fundamental goal in artificial intelligence, continual learning grapples with an enduring challenge—catastrophic forgetting. A proficient model must exhibit adaptability for new data assimilation and robustness to retain existing knowledge. Class-incremental learning (CIL) aids the gradual integration of knowledge from newly introduced classes, forming a universal classifier. However, directly training the model with fresh class instances triggers a problem—forgetting distinguishing features of prior classes, causing a performance decline. Addressing such issues in machine learning, this survey aims to delineate significant challenges, outcomes, and recent advancements, including our contributions to CIL techniques, especially in image classification.

The Van der Pol oscillator is a chaotic non-linear system. Small perturbations in initial conditions may result in wildly different trajectories. Because of its chaotic nature, controlling, or forcing, the behavior of a Van der Pol oscillator is difficult to achieve through traditional adaptive control methods. Connecting two Van der Pol oscillators together where the output of one oscillator, the driver, drives the behavior of its partner, the responder, is a proven technique for controlling the Van der Pol oscillator. Deterministic AI (DAI) is an adaptive feedback control method that leverages the known physics of the Van der Pol system to learn optimal system parameters for the forcing function. We assessed the performance of DAI employing three different online parameter estimation algorithms. Our evaluation criteria include mean absolute error (MAE) between the target trajectory and the response oscillator trajectory over time. RLS with exponential forgetting (RLS-EF) had the lowest MAE overall, with a 2.46% reduction in error. However, another method was notable. Least Mean Squares with normalized gradient adaptation (LMS-NG) had worse initial error in the first 10% of the simulation, but after that point had consistently better performance. We found that over the last 90% of the simulation, DAI with LMS-NG had a 48.7% reduction in MAE compared to feedforward alone.