Adaptive Mapping Algorithm Research Articles

Universal Adaptive Fault-Tolerant Control of a Multicopter UAV

Abstract This paper presents a universal adaptive fault-tolerant control (FTC) design for multicopter unmanned aerial vehicles (UAVs). The proposed architecture consists of a two-loop control structure: a fault-tolerant controller generates normalized virtual control inputs to track the desired trajectory subject to actuator faults, and an adaptive augmentation controller deals with system uncertainties and also balances the design requirements for specific platform. The FTC approach is based on gain-scheduling control in the framework of structured H∞ synthesis. In order to implement the overall control system on most types of multicopter UAVs, an adaptive mapping algorithm is proposed. High fidelity simulations and experimental results, performed on various multicopters with different payload and configuration, show the effectiveness and robustness of the proposed approach in accommodating different levels of actuator degradation including total failures of the rotors as well as unknown mass and inertia, all using a single controller with fixed coefficients.

Codem: software/hardware codesign for embedded multicore systems supporting hardware services

Efficient software/hardware codesign is posing significant challenges to embedded systems. This paper proposes Codem, a software/hardware codesign flow for embedded systems, which models both processors and Intellectual Property (IP) cores as services. Tasks are regarded as abstract instructions which can be scheduled to IP cores for parallel execution automatically. In order to guide the hardware implementations of the hot spot functions, this paper incorporates a novel hot spot-based profiling technique to observe the hot spot functions while the application is being simulated. Furthermore, based on the hot spot of various applications, an adaptive mapping algorithm is presented to partition the application into multiple software/hardware tasks. We test the profiling-based design flow with classic Sort applications. Experimental results demonstrate that Codem can efficiently help researchers to identify the hot spots, and also outline a new direction to combine profiling techniques with state-of-the-art reconfigurable computing platforms for specific task acceleration.

Self-Organizing Fuzzy Haptic Teleoperation of Mobile Robot Using Sparse Sonar Data

Mobile robot teleoperation has been used in many areas of industrial automation, such as explosives disposal, nuclear waste manipulation, freight handling, or transportation. Here, the commonly provided audio-visual feedback often resulted in an inadequate perception of the remote environment. Haptic augmentation was shown to improve and positively enhance the control of the mobile robot. This paper presents novel self-organizing fuzzy adaptive mapping algorithm (SOFAMap) for a haptic teleoperation of mobile robots. The SOFAMap algorithm was specifically developed for a mobile robot with a rotational sonar sensory system, constituting an alternative to a traditionally used multisonar array. The main contributions of this paper are the following: 1) development of a specific self-organizing environment mapping structure inspired by the growing neural gas algorithm; 2) incorporating a fuzzy controller into the algorithm to adapt to robot's motion; and 3) resolving typical issues such as sensor noise, communication time delay, and low sampling rate. The experimental testing was performed in both virtual environment and on a real robotic platform, consisting of a Lego NXT mobile robot and a Novint Falcon 3-DOF haptic interface. The results showed that a high-fidelity haptic feedback can be successfully generated using a simpler and more affordable rotational sonar sensory system, as opposed to the typical multisonar array. Furthermore, it was demonstrated that the SOFAMap algorithm improves the operator's awareness of unstructured environments, making it applicable to wide range of mobile robot teleoperation systems.

Introducing probabilistic adaptive mapping developmental genetic programming with redundant mappings

Developmental Genetic Programming (DGP) algorithms have explicitly required the search space for a problem to be divided into genotypes and corresponding phenotypes. The two search spaces are often connected with a genotype-phenotype mapping (GPM) intended to model the biological genetic code, where current implementations of this concept involve evolution of the mappings along with evolution of the genotype solutions. This work presents the Probabilistic Adaptive Mapping DGP (PAM DGP), a new developmental implementation that involves research contributions in the areas of GPMs and coevolution. The algorithm component of PAM DGP is demonstrated to overcome coevolutionary performance problems that are identified and empirically benchmarked against the latest competing algorithm that adapts similar GPMs. An adaptive redundant mapping encoding is then incorporated into PAM DGP for further performance enhancement. PAM DGP with two mapping types are compared to the competing Adaptive Mapping algorithm and Traditional GP in two medical classification domains, where PAM DGP with redundant encodings is found to provide superior fitness performance over the other algorithms through it's ability to explicitly decrease the size of the function set during evolution.

Connectionist modeling for arm kinematics using visual information

The self-organizing adaptive map algorithm is adopted to learn all possible postures for an artificial arm of arbitrary configuration placed in a three-dimensional workspace. Arm postures are represented through their projections onto a set of image planes. Based on the link orientation and link length extracted from these images, a topological state space Q is generated. Arm kinematics is expressed as a transformation of topological hypersurfaces, the intersections of which represents the multiple postures of the arm in the workspace for a given end effector position. The self-organizing feature map learns how the topological hypersurfaces transform in the state space during arbitrary movements of the arm in the workspace. During the learning phase, the neural network generates clusters of neurons, each neuron being responsible for reproducing an arm posture in the workspace. The neural clusters map the hypersurfaces' intersection in the topological Q-space to any position of the arm gripper in the workspace. Simulations for planar and nonplanar multiple degrees of freedom arms are presented.