Earticle

In this application of artificial intelligence to a real-world problem, the constrained scheduling of employee resourcing for a mall type shop is solved by means of a genetic algorithm. hromosomes encode a one-week schedule and a constraint matrix handles all requirements for the population. The genetic operators are purposely designed to preserve all constraints and the objective function assures an imposed coverage, this is for people on both sections of the mall. The results demonstrate that the genetic algorithm approach can provide acceptable solutions to this type of employee scheduling problem with constrains.

The exposure of integrated electronic components and systems to ionizing radiation may lead to minor deterioration in performance or catastrophic system failure, depending on the level of radiation (as a function of altitude). Mitigation of the radiation-induced hazards is of a major concern for space applications, since electronic components are expected to function without failure for an extended period of time, under extreme operating conditions. The work presented in this paper examines the effect of radiation on electronic components through analytical modeling of parameters, some of which are design-dependent and others are process-dependent. Examinations of the inter-dependency of these parameters would aid identifying possible solutions to the radiation tolerance problem.

Self configuring VLSI technology architectures offer a new environment for creating novel security functions. Two such functions for physical security architectures are proposed to be generated autonomously as unknown/secret internal functions. A cell-based FPGA technology architecture is deployed for generating two classes of self-constructed one-way physical secret functions, one representing a hash function and the other a ciphering function. The Hash function is a non-invertible mapping, where the cipher function should be invertible. The two sample architectures of the functions are inspired from the programmable cell structure of the selected FPGA technology. As the functions are internally created, their mapping structures can be kept completely secret and even unknown to anybody. Such units could be efficiently deployed for a novel physical security even when nothing is known about their exact architecture and mapping functions. Several new attractive application scenarios are demonstrated including a type of zero-knowledge proof of identity and clone-resistant physical units as well as secured dependency functions. It is also shown that such security mechanisms can be kept operational for some useful applications even if the secret-unknown functions are allowed to evolve and develop additional time-dependent and individual properties. Such security functions became recently possible after self-configuring VLSI architectures are available as a part of real microelectronic systems.

Robot Security is becoming more and more a serious issue for many modern applications. Robot Security matters are still not intensively addressed in the published literature. The goal of this paper is to explore possible identification and security mechanisms which fit to robot technologies and their operating environment. To secure transactions between robots deployed in open service, robots need first to be securely identified “as born persons” with unique provable identities. Robots rolling from a production line are assumed to be equal objects; therefore the first necessary action is to personalize robots and give them unclonable identities. A sort of “Electronic mutation” technology was specified and proposed to create a non-reversible and non-repeatable robot identity, which is at the same time securely provable [6]. The identity exhibits properties similar to those of human DNA. The resulting cloneresistant or (unclonable) identity is adapted and proposed to be embedded in a robot environment. The goal is also to diffuse the identity traces possibly into all robot activities similarly as the human DNA do throughout the whole body of a biological creature. The identity proposed is made traceable through cryptographic signatures linked to relevant robot mechanical and electronic activities as a step towards “mechatronic security”. The work is also aiming to stimulate discussions on robot security issues or in general the question of “mechatronic security”.

This paper presents a discussion concerning EEG signals compression using the basis pursuit (BP) approach applied for several overcomplete wavelet dictionaries. The compression is based on an “optimal” superposition of dictionary elements, by minimizing the l1 norm of the error. The best results have been obtained with the Daubechies10 dictionary.