Autonomous Driving Challenge: To Infer the Property of a Dynamic Object Based on Its Motion Pattern

Abstract

In autonomous driving applications a critical challenge is to identify the action to take to avoid an obstacle on a collision course. For example, when a heavy object is suddenly encountered it is critical to stop the vehicle or change the lane even if it causes other traffic disruptions. However, there are situations when it is preferable to collide with the object rather than take an action that would result in a much more serious accident than collision with the object. For example, a heavy object which falls from a truck should be avoided whereas a bouncing ball or a soft target such as a foam box need not be.We present a novel method to discriminate between the motion characteristics of these types of objects based on their physical properties such as bounciness, elasticity, etc.In this preliminary work, we use recurrent neural network with LSTM (Long Short Term Memory) cells to train a classifier to classify objects based on their motion trajectories. We test the algorithm on synthetic data, and, as a proof of concept, demonstrate its effectiveness on a limited set of real-world data.

Machine Learning for Smart Environments in B5G Networks: Connectivity and QoS

“…Works in [131][132][133][134] discuss IoT device's connectivity and wireless communication using ANN. Here, ANN plays a vital role in enhancing the driver behavior modeling [131], classifying the objectives [133,134], and predicting the speed of mobility [132].…”

Self-Supervised Learning for Autonomous Vehicles Perception: A Conciliation Between Analytical and Learning Methods

“…In turn, these tracking approaches provide the tools to collect motion pattern information of surrounding dynamic obstacles, such that this information may help to classify obstacles depending on their dynamic properties [53].…”

$DeepMotions$ : A Deep Learning System for Path Prediction Using Similar Motions

“…Other attempts tried predicting future paths based on recurrent neural networks notions. notions [7], [9], [11], [27], [28], [31], [51], [53]. In [27], authors propose a novel method named T-CONV which models trajectories as two-dimensional images and adopts multi-layer convolutional neural networks to combine multi-scale trajectory patterns to obtain highly precise predictions and extract the areas with distinct influence on the ultimate prediction.…”

“…In [53], the authors introduce a neural network model named RA-LSTM which combines the road-aware features to predict the future trajectory path. Several techniques employing LSTM (Long Short-Term Memory) cells for predicting future trajectories have been introduced in the literature [7], [11], [28], [31], [51].…”