Learning traversability models for autonomous mobile vehicles

Autonomous mobile robots need to adapt their behavior to the terrain over which they drive, and to predict the traversability of the terrain so that they can effectively plan their paths. Such robots usually make use of a set of sensors to investigate the terrain around them and build up an internal representation that enables them to navigate. This paper addresses the question of how to use sensor data to learn properties of the environment and use this knowledge to predict which regions of the environment are traversable. The approach makes use of sensed information from range sensors (stereo or ladar), color cameras, and the vehicle’s navigation sensors. Models of terrain regions are learned from subsets of pixels that are selected by projection into a local occupancy grid. The models include color and texture as well as traversability information obtained from an analysis of the range data associated with the pixels. The models are learned without supervision, deriving their properties from the geometry and the appearance of the scene. The models are used to classify color images and assign traversability costs to regions. The classification does not use the range or position information, but only color images. Traversability determined during the model-building phase is stored in the models. This enables classification of regions beyond the range of stereo or ladar using the information in the color images. The paper describes how the models are constructed and maintained, how they are used to classify image regions, and how the system adapts to changing environments. Examples are shown from the implementation of this algorithm in the DARPA Learning Applied to Ground Robots (LAGR) program, and an evaluation of the algorithm against human-provided ground truth is presented.

Design and dynamics of an innovative micro gyroscope against coupling effects

This paper presents an innovative micro gyroscope design. The proposed tri-axis gyroscope possesses the capability of detecting three-dimensional angular motions. The motion of each sensing element is, by elaborate mechanical design, restricted to move in orthogonal direction to each other such that the measurements by high-resolution capacitors with signal processing circuits are decoupled and precisely represent, to some extent, angular velocity components in three axes. The drive electrode comb is used to constantly vibrate the proof mass in tangential direction by sinusoidal voltage. The signal bandwidth is increased by distributed translational proof masses, placed ninety degree apart from each other. Each individual proof mass is designed to move solely in radial direction so that superior mode matching can be achieved. In order to ensure better repeatability and more reliability, the suspension flexures and damping effects are studied such that stress of the proposed micro gyroscope is reduced but the span of angular displacements is increased. Owing to the complicated geometry of the suspension flexures, finite element method (FEM) is employed to obtain more exact stiffness values and compared with theoretical analysis. The dynamic model of the proposed gyroscope is established to include non-linear terms and embedded mechanical constraints. The entire micro device can be produced merely by surface fabrication such that the mass production cost can be considered at the design stage, while the resolution, bandwidth and decoupling capability of tri-axis detection are enhanced.     

Utilize bootstrap in small data set learning for pilot run modeling of manufacturing systems

If the production process, production equipment, or material changes, it becomes necessary to execute pilot runs before mass production in manufacturing systems. Using the limited data obtained from pilot runs to shorten the lead time to predict future production is this worthy of study. Although, artificial neural networks are widely utilized to extract management knowledge from acquired data, sufficient training data is the fundamental assumption. Unfortunately, this is often not achievable for pilot runs because there are few data obtained during trial stages and theoretically this means that the knowledge obtained is fragile. The purpose of this research is to utilize bootstrap to generate virtual samples to fill the information gaps of sparse data. The results of this research indicate that the prediction error rate can be significantly decreased by applying the proposed method to a very small data set.     

Hybrid high order neural networks

Neural networks (NNs) represent a familiar artificial intelligence approach widely applied in many fields and to a wide range of issues. The back propagation network (BPN) is one of the most well-known NNs, comprising multilayer perceptrons (MLPs) with an error back propagation learning algorithm. BPN typically employs associate multiplicative weightings for layer connections. For single connections, BPN combines neuron inputs linearly to neuron outputs. In this study, the author develops and embeds high order connections (exponent multipliers) into the BPN. The resultant proposed hybrid high order neural network (HHONN) is intended to be applicable to both linear and high order connections. HHONN allows an additional connection type for BPN, which permits BPN to adapt to different scenarios. In this paper, learning equations for both weighting and high order connections are introduced in their general forms. A feedforward neural network with a topology of two hidden layers and one high order connection was developed and studied to confirm the improved performance of developed HHONN models. Case studies, including two basic tests (a function approximation and the TC problem) and squat wall strength learning, were used to verify HHONN performance. Results showed that, when the high order connection was employed anywhere except the eventual connection, HHONN delivered better results than achievable using traditional BPN. Such results show that HHONN successfully introduces high order connections into BPN.     

The evaluation of hot-working characteristics of cobalt-based implant alloys

The hot-working characteristics of wrought Co-Ni-Cr-Mo implant alloy during ingot-to-billet conversion were evaluated using a Gleeble-2000A simulator. The hot tensile test at 700–1 320 °C was used to determine the optimum hot-working parameters at a strain rate equivalent to that of conventional press forging to ensure acceptable hot workability. Hot ductility and deformation resistance as a function of temperature can be clearly established. The fracture surfaces of the tensile specimens were examined to correlate them with the hot tensile ductility values at various temperatures. The poor ductility at temperatures above 1300 °C was attributed to the incipient melting of grain boundaries. The effect of temperature and strain rate on the flow-stress behaviour and microstructures were investigated by uniaxial compression testing in the temperature range 900–1200 °C and strain rate, , range of 0.01–10s^–1. The strain-hardening and steady-state behaviour were described from the measured true stress-true strain curves.     

Lengthening kinetics of (01\bar 10) γ-TiH precipitates in α-Ti in the temperature range of 25 °c to 80 °c

Formation of y hydrides in α titanium involves a hcp → fct lattice transformation with hydrogen as an interstitial diffusing species. Since the crystallography of this reaction resembles that of a martensitic transformation and the hydrides form at temperatures where both titanium and hydrogen atoms may have negligible mobility, many investigators have concluded that γ hydrides form by a martensitic mechanism. In this study, the lengthening kinetics of γ hydrides were determined using conventional transmission electron microscopy (TEM). The length of the hydrides was measured as a function of time, temperature, and hydrogen concentration in samples isothermally aged in the range of 25 °C to 80 °C. These measurements were compared with the Zener-Hillert model of volume-diffusion-controlled precipitate growth. The experimental results indicate that y hydride growth is diffusionally controlled at 25 °C and possibly interfacially controlled at aging temperatures of 50 °C and higher. These and other aspects of the transformation are discussed.     

Open Innovation and Team Leaders’ Innovation Traits

Open innovation has transformed internal novel resource allocation and facilitated diverse interorganizational cooperation models. Previous studies have investigated open innovation archetypes and their formation. However, most researchers have neglected the crucial role that team leaders play in integrating internal and external resources and connecting creative ideas in open innovation. This study investigated the influence of open innovation archetypes and team leaders’ innovation traits on team performance. On the basis of the literature review, the analytical hierarchy process was applied to construct a framework for assessing and selecting team leaders in open innovation. The analysis was completed based on samples from 45 valid respondents (88% return rate) in Taiwan’s high-tech industry. The study results suggested that innovation knowledge is the most crucial innovation criterion for a leader, particularly core knowledge and advanced knowledge. Among the various sub-criteria, extrinsic motivation was the second most important sub-criterion. Leaders with strong innovation motivations are suitable to lead inbound open innovation projects.     

A new carbon structure in annealed film coatings of the carbon–lead system

Carbon–lead solid solutions coexisting with amorphous carbon have been obtained for the first time in a film coating deposited by ion-plasma sputtering. During subsequent vacuum annealing of carbon–lead films containing more than 68.5 at % Pb, this element almost completely evaporates to leave an amorphous carbon coating on a substrate. During annealing at 1100°C, this amorphous carbon crystallizes into a new hexagonal lattice with unit cell parameters a = 0.7603 nm and c = 0.8168 nm. Characteristic X-ray diffraction data for the identification of this phase are determined.

     

Study on the Impurity Effect in the Realization of Silver Fixed Point

The application of a thermal analysis model to estimate the temperature depression from the ideal fixed-point temperature is important, especially when the chemical analysis of the sample in a cell is insufficient or the cell might have been contaminated during fabrication. This study extends previous work, on thermal analysis with the tin point, to an investigation of the impurity dependence of the silver-point temperature. Close agreement was found between the temperature depression (\(-0.36\) mK) inferred from the thermal analysis of the measured fixed-point plateau and the temperature depression (\(-0.32\) mK) inferred using the sum of individual estimates (SIE) method with an impurity analysis based on glow discharge mass spectrometry. Additionally, the results of the thermal analysis manifest no significant dependence on the rate of solidification, and the scatter of observed gradients was within 0.36 mK among five plateaux with different temperature settings of the furnace. Although the results support the application of both the SIE method and thermal analysis for the silver point, further experiments with cell-to-cell comparisons linked to thermal analysis, a study of the thermometer-furnace systematic effects, the oxygen effect, and the locus of the freezing plateau should be investigated to reach a firm conclusion.     

The lateral load-deformation response of RC walls with a consideration of the height–length ratio

Reinforced concrete (RC) walls play an important role in resisting earthquakes, so understanding the lateral load-deformation response of an RC wall subjected to an axial load, shear, and moment is essential to nonlinear structural analysis. In this study, the moment-bending deformation response of an RC wall under an axial load and moment is obtained from moment–curvature analysis using the wall’s proper plastic hinge length. Furthermore, modified compression-field theory, adjusted according to the height–length ratio of the wall, is used to calculate the shear–shear deformation response of the wall under an axial load and shear. By integrating the moment-bending deformation and the shear–shear deformation responses, the lateral load-deformation response of the RC wall under axial load–moment–shear interaction can be reasonably determined. To confirm the reliability of the proposed method, experimental results for 67 RC walls are compared with analysis results from the proposed method. The statistical results show that the proposed method accurately predicts the lateral ultimate load but somewhat underestimates the lateral ultimate deformation. Finally, this paper gives an example using an equivalent column to simulate an RC wall using SAP2000 pushover analysis.