A new semi-supervised PSVM classifier

Zhou, Anwa; Zhu, Zhibin; Fan, Hao

doi:10.1016/j.amc.2012.10.037

Cited by 4 publications

(4 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When the Lagrange multipliers are set to zero, that is, ε * = 0 and/or ε * = 0, the control τ * can be calculated (including the solution at the boundaries) by solving (45). Otherwise, when the Lagrange multipliers are non-zero, that is, ε * = 0 and/or ε * = 0, the control τ * has to be equal to the boundary values, τ max or τ min , based on the KKT conditions.…”

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

“…The KKT conditions [44] are the necessary conditions for the local minima and global optimal solution, and the algorithm is independent of the initial conditions. The KKT necessary and sufficient optimality conditions [45] are obtained by setting the gradients of the Lagrangian L equal to zero. Then, the optimal solution, τ * , is obtained by solving (see equation at the bottom of the page)…”

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

\begin{matrix} right left1em4pt \\ {(|, \frac{\partial bold-italicL}{\partial τ})}_{\begin{matrix} τ = τ^{*} \\ \underline{ε} = {\underline{ε}}^{*} \\ \overset{false¯}{ε} = {\bar{ε}}^{*} \end{matrix}} & = 2 D^{T} W_{u}^{T} W_{u} (bold-italicD {bold-italicτ}^{*} - bold-italicu) \\ + ε \frac{diag (((\partial {bold-italicδ}_{normalo i} ({bold-italicτ}_{i}^{*})) / (\partial {bold-italicτ}_{i})) {bold-italicς}_{normalo i} ({bold-italicτ}_{i}^{*}) - δ_{normalo i} ({bold-italicτ}_{i}^{*}) ((\partial {bold-italicς}_{normalo i} (τ_{i}^{*})) / (\partial τ_{i})))}{{bold-italicς}_{o}^{2} ({bold-italicτ}_{i}^{*})} - {\underline{ε}}^{*} + {\overset{false¯}{ε}}^{*} = 0 \end{matrix}

…”

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

See 2 more Smart Citations

Finite‐time fault tolerant attitude control for over‐activated spacecraft subject to actuator misalignment and faults

Zhang

Friswell

2013

IET Control Theory & Applications

View full text Add to dashboard Buy / Rent full text

A finite-time attitude compensation control scheme is developed for an over-activated rigid spacecraft subject to actuator faults, misalignment, external disturbances and uncertain inertia parameters. The controller is synthesised based on the sliding mode control theory, and guarantees the finite-time reachability of the system states. A sufficient condition for the controller to accommodate misalignment and faults of actuator is presented. An optimised control allocation algorithm based on the Karush-Kuhn-Tucker condition is then applied to distribute the commanded control to each actuator, and optimise the consumption of energy. Numerical simulation results are presented that highlight the performance benefits of the control law.

show abstract

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

\begin{matrix} right left1em4pt \\ {(|, \frac{\partial bold-italicL}{\partial τ})}_{\begin{matrix} τ = τ^{*} \\ \underline{ε} = {\underline{ε}}^{*} \\ \overset{false¯}{ε} = {\bar{ε}}^{*} \end{matrix}} & = 2 D^{T} W_{u}^{T} W_{u} (bold-italicD {bold-italicτ}^{*} - bold-italicu) \\ + ε \frac{diag (((\partial {bold-italicδ}_{normalo i} ({bold-italicτ}_{i}^{*})) / (\partial {bold-italicτ}_{i})) {bold-italicς}_{normalo i} ({bold-italicτ}_{i}^{*}) - δ_{normalo i} ({bold-italicτ}_{i}^{*}) ((\partial {bold-italicς}_{normalo i} (τ_{i}^{*})) / (\partial τ_{i})))}{{bold-italicς}_{o}^{2} ({bold-italicτ}_{i}^{*})} - {\underline{ε}}^{*} + {\overset{false¯}{ε}}^{*} = 0 \end{matrix}

…”

Section: Control Allocation With Energy Optimisationmentioning

confidence: 99%

See 1 more Smart Citation

Finite‐time fault tolerant attitude control for over‐activated spacecraft subject to actuator misalignment and faults

Zhang

Friswell

2013

IET Control Theory & Applications

View full text Add to dashboard Buy / Rent full text

show abstract

“…Naseem et al gave a face recognition method as Linear Regression based Classification (LRC) [10]. Zhou et al [11] developed a semi-supervised Proximal Support Vector Machine (PSVM) classifier.…”

Section: Introductionmentioning

confidence: 99%