Pedestrian Intention and Pose Prediction through Dynamical Models and Behaviour Classification

Quintero, R.; Parra, I.; Llorca, David Fernández; Sotelo, Miguel Ángel

doi:10.1109/itsc.2015.22

Cited by 38 publications

(26 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Keller and Gavrila [12] do this by predicting the dense optical flow of the VRU's image, using Gaussian Process Dynamic Models (GPDM). Quintero et al [5] use GPDM on a 3D joint model that is extracted from the visual and depth image of the pedestrian. VRU-related context cues also include whether the VRU is aware of his or her surroundings.…”

Section: Previous Workmentioning

confidence: 99%

“…Path prediction also requires a dynamic model to propagate the positional and contextual information into the future. One approach is to create separate models for different kinds of behavior and predict which behavior is currently most likely [4], [5], which makes the entire model more interpretable to users. However, due to the non-linearity in VRU path prediction some research has pursued RNNs to get an improved prediction accuracy, such as Long Short Term Memory networks (LSTMs) [10], [11].…”

Section: Previous Workmentioning

confidence: 99%

“…They can rapidly switch between various motion modes, such as walking or standing. Separate models are often learned for each of the identified modes [2], [5]. The various contextual cues need to be integrated into a predictive model to assess which mode the VRU will switch to, and when the VRU will do so, e.g.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Context-based cyclist path prediction using Recurrent Neural Networks

Pool

Kooij

Gavrila

2019

2019 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

This paper proposes a Recurrent Neural Network (RNN) for cyclist path prediction to learn the effect of contextual cues on the behavior directly in an end-to-end approach, removing the need for any annotations. The proposed RNN incorporates three distinct contextual cues: one related to actions of the cyclist, one related to the location of the cyclist on the road, and one related to the interaction between the cyclist and the egovehicle. The RNN predicts a Gaussian distribution over the future position of the cyclist one second into the future with a higher accuracy, compared to a current state-of-the-art model that is based on dynamic mode annotations, where our model attains an average prediction error of 33 cm one second into the future.

show abstract

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

See 1 more Smart Citation

Context-based cyclist path prediction using Recurrent Neural Networks

Pool

Kooij

Gavrila

2019

2019 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

show abstract

“…In particular, [3] address this problem by adding pedestrian's head orientation as an intention feature to assess the criticality of situations. Other features that may be used are the human pose [17], [18] and optical flow [16] to determine the state of a pedestrian.…”

Section: Related Workmentioning

confidence: 99%

Composable Q- Functions for Pedestrian Car Interactions

Muench

Gavrila

2019

2019 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

We propose a novel algorithm that predicts the interaction of pedestrians with cars within a Markov Decision Process framework. It leverages the fact that Q-functions may be composed in the maximum-entropy framework, thus the solutions of two sub-tasks may be combined to approximate the full interaction problem. Sub-task one is the interactionfree navigation of a pedestrian in an urban environment and sub-task two is the interaction with an approaching car (deceleration, waiting etc.) without accounting for the environmental context (e.g. street layout). We propose a regularization scheme motivated by the soft-Bellman-equations and illustrate its necessity. We then analyze the properties of the algorithm in detail with a toy model. We find that as long as the interactionfree sub-task is modelled well with a Q-function, we can learn a representation of the interaction between a pedestrian and a car.

show abstract

“…Quintero et al [8] used Balanced Gaussian Process Dynamical Models and a naïve-Bayes classifier for intention and pose prediction of pedestrians based on 3D joint positions. This approach was extended by a Hidden Markov Model in [5].…”

Section: B Related Workmentioning

confidence: 99%

Early Start Intention Detection of Cyclists Using Motion History Images and a Deep Residual Network

Zernetsch

Kress

Sick

et al. 2018

2018 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

In this article, we present a novel approach to detect starting motions of cyclists in real world traffic scenarios based on Motion History Images (MHIs). The method uses a deep Convolutional Neural Network (CNN) with a residual network architecture (ResNet), which is commonly used in image classification and detection tasks. By combining MHIs with a ResNet classifier and performing a frame by frame classification of the MHIs, we are able to detect starting motions in image sequences. The detection is performed using a wide angle stereo camera system at an urban intersection. We compare our algorithm to an existing method to detect movement transitions of pedestrians that uses MHIs in combination with a Histograms of Oriented Gradients (HOG) like descriptor and a Support Vector Machine (SVM), which we adapted to cyclists. To train and evaluate the methods a dataset containing MHIs of 394 cyclist starting motions was created. The results show that both methods can be used to detect starting motions of cyclists. Using the SVM approach, we were able to safely detect starting motions 0.506 s on average after the bicycle starts moving with an F 1 -score of 97.7%. The ResNet approach achieved an F 1score of 100% at an average detection time of 0.144 s. The ResNet approach outperformed the SVM approach in both robustness against false positive detections and detection time.S. Zernetsch, V. Kress and K. Doll are with the Faculty of Engineering, University of Applied Sciences Aschaffenburg,

show abstract

Pedestrian Intention and Pose Prediction through Dynamical Models and Behaviour Classification

Cited by 38 publications

References 9 publications

Context-based cyclist path prediction using Recurrent Neural Networks

Context-based cyclist path prediction using Recurrent Neural Networks

Composable Q- Functions for Pedestrian Car Interactions

Early Start Intention Detection of Cyclists Using Motion History Images and a Deep Residual Network

Contact Info

Product

Resources

About