Comparing Methods of DC Motor Control for UUVs

Abstract

Adaptive and learning methods are proposed and compared to control DC motors actuating control surfaces of unmanned underwater vehicles. One type of adaption method referred to as model-following is based on algebraic design, and it is analyzed in conjunction with parameter estimation methods such as recursive least squares, extended least squares, and batch least squares. Another approach referred to as deterministic artificial intelligence uses the process dynamics defined by physics to control output to track a necessarily specified autonomous trajectory (sinusoidal versions implemented here). In addition, one instantiation of deterministic artificial intelligence uses 2-norm optimal feedback learning of parameters to modify the control signal, while another instantiation is presented with proportional plus derivative adaption. Model-following and deterministic artificial intelligence are simulated, and respective performance metrics for transient response and input tracking are evaluated and compared. Deterministic artificial intelligence outperformed the model-following approach in minimal peak transient value by a percent range of approximately 2–70%, but model-following achieved at least 29% less error in input tracking than deterministic artificial intelligence. This result is surprising and not in accordance with the recently published literature, and the explanation of the difference is theorized to be efficacy with discretized implementations.

Impacts of Discretization and Numerical Propagation on the Ability to Follow Challenging Square Wave Commands

“…Zhang et al [17] illustrated fault-tolerance, while Apoorva et al [18] revealed loss reduction and Flieh et al demonstrated loss minimization [19] and dead-beat control [20] in addition to self-sensing [21], the precursor to using the physics-based dynamics for virtual sensing [22] following the illustration of optimality in [23] and self-sensing [24] specifically applied to DC motors. Despite stochastic learning methods still holding some interest [25] applied to motor control, this manuscript continues the investigation of deterministic learning approaches [26] following Shah's recommendations [27]. Specifically, [26] illustrated a marked improvement in tracking performance, while Shah's attempt in [27] to duplicate the results revealed a strong correlation to performance improvement and system discretization and speed of computation.…”

“…This manuscript proposes a preferred instantiation of adaptive and learning systems [26,27] by evaluating the efficacy of motor control techniques based on iterated computational rates and system discretization. The materials and methods in Section 2 first describe model discretization and then introduces the two compared: one adaptive and one learning each with interconnected lineage of research in the literature.…”

“…The momentum continued, and the nonlinear output regulation has been further explored by numerous authors including Cheng, Tarn, and Spurgeon [37], Khalil [38], and Wang and Huang [39] across autonomous and nonautonomous systems. The lineage emphasized in this manuscript stems from a heritage in vehicle guidance and control techniques [8][9][10][11][12][13][14][15] extended to apply to motor controllers [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35] that generate vehicle motion. Vehicle maneuvering is controlled by the actuator fins displayed in Figure 3b generating navigation as displayed in Figure 3a.…”

Synthetic Data Augmentation and Deep Learning for the Fault Diagnosis of Rotating Machines

“…In practice, there exists an issue of data imbalance where there is sufficient data from the healthy state, while very limited data is available from the faulty states of the system. Contrary to the stochastic machine learning and deep learning models, deterministic artificial intelligence (DAI) takes into account the first principles (i.e., underlying physics of the problem) whenever available [23,24].…”

The WL_PCR: A Planning for Ground-to-Pole Transition of Wheeled-Legged Pole-Climbing Robots

“…There are also some methods to consider the influence of uncertainty when designing the reference orbit in advance [20][21][22]. For example, in the process of motion control, the uncertain performance of parameters was used to establish the optimal control scheme [20]. Desensitization optimal control [22] modifies the nominal optimal trajectory to reduce its sensitivity to uncertain parameters.…”