Specifying Object Behavior

A Lightweight Motional Object Behavior Prediction System Harnessing Deep Learning Technology for Embedded ADAS Applications

Electronics ◽

10.3390/electronics10060692 ◽

2021 ◽

Vol 10 (6) ◽

pp. 692

Author(s):

Wen-Chia Tsai ◽

Jhih-Sheng Lai ◽

Kuan-Chou Chen ◽

Vinay M.Shivanna ◽

Jiun-In Guo

Keyword(s):

Embedded System ◽

Moving Objects ◽

Object Localization ◽

Crossing Over ◽

Learning Technology ◽

Prediction System ◽

Behavior Prediction ◽

Detection Model ◽

Pedestrian Crossing ◽

Object Behavior

This paper proposes a lightweight moving object prediction system to detect and recognize pedestrian crossings, vehicles cutting-in, and vehicles ahead applying emergency brakes based on a 3D Convolution network for behavior prediction. The proposed design significantly improves the performance of the conventional 3D convolution network (C3D) adapted to predict the behaviors employing behavior recognition network capable of performing object localization, which is pivotal in detecting the numerous moving objects’ behaviors, combining and verifying the detected objects with the results of the YOLO v3 detection model with that of the proposed C3D model. Since the proposed system is a lightweight CNN model requiring far lesser parameters, it can be efficiently realized on an embedded system for real-time applications. The proposed lightweight C3D model achieves 10 frames per second (FPS) on a NVIDIA Jetson AGX Xavier and yields over 92.8% accuracy in recognizing pedestrian crossing, over 94.3% accuracy in detecting vehicle cutting-in behavior, and over 95% accuracy for vehicles applying emergency brakes.

Download Full-text

Noncontacting Vibration Measurement Using a Massless Interferometric Accelerometer

Volume 3C: 15th Biennial Conference on Mechanical Vibration and Noise — Vibration Control, Analysis, and Identification ◽

10.1115/detc1995-0721 ◽

1995 ◽

Author(s):

Keyu Li

Keyword(s):

Elevated Temperatures ◽

Laser Light ◽

Vibration Measurement ◽

Object Surface ◽

Out Of Plane ◽

Plates And Shells ◽

Fringe Patterns ◽

Hostile Environments ◽

Derivatives Of ◽

Object Behavior

Abstract An interferometric strain measurement technique is extended to vibration measurements. The technique is based on two micro-indentations placed on an object surface using a combination of diffraction and interference of laser light. Relative displacements between the two indentations and derivatives of in-plane and out-of-plane vibrational displacements are measured by analyzing the phase shift of the interference fringe patterns. The technique can be used to study bending stress and deflection problems in vibrational beams, plates and shells. The displacement derivatives are measured in real time, from which time derivatives or the velocity and acceleration of the displacement derivative as well as vibrational frequency can be determined. The technique has advantages over an accelerometer in that it is noncontacting and does not require attachment of the transducer to the object which could alter the object behavior. In addition, it has many desirable features such as being extremely compact, massless, and applicable to hostile environments such as those associated with production and elevated temperatures.

Download Full-text

FORMALIZATION OF OBJECT BEHAVIOR AND INTERACTIONS FROM UML MODELS

International Journal of Software Engineering and Knowledge Engineering ◽

10.1142/s021819400100075x ◽

2001 ◽

Vol 11 (06) ◽

pp. 643-673 ◽

Cited By ~ 31

Author(s):

JOHN ANIL SALDHANA ◽

SOL M. SHATZ ◽

ZHAOXIA HU

Keyword(s):

Petri Net ◽

System Level ◽

Simulation Techniques ◽

Industry Standard ◽

Automated Processing ◽

Language Needs ◽

Object Models ◽

Collaboration Diagram ◽

Object Behavior

UML, being the industry standard as a common OO modeling language, needs a well-defined semantic base for its notation. Formalization of the graphical notation enables automated processing and analysis tasks. This paper describes a methodology for synthesis of a Petri net model from UML diagrams. The approach is based on deriving Object Net Models from UML statechart diagrams and connecting these object models based on UML collaboration diagram information. The resulting system-level Petri net model can be used as a foundation for formal Petri net analysis and simulation techniques. The methodology is illustrated on some small examples and a larger case study. The case study reveals some unexpected invalid system-state situations.

Download Full-text