-
FlexLoc: Conditional Neural Networks for Zero-Shot Sensor Perspective Invariance in Object Localization with Distributed Multimodal Sensors
Authors:
Jason Wu,
Ziqi Wang,
Xiaomin Ouyang,
Ho Lyun Jeong,
Colin Samplawski,
Lance Kaplan,
Benjamin Marlin,
Mani Srivastava
Abstract:
Localization is a critical technology for various applications ranging from navigation and surveillance to assisted living. Localization systems typically fuse information from sensors viewing the scene from different perspectives to estimate the target location while also employing multiple modalities for enhanced robustness and accuracy. Recently, such systems have employed end-to-end deep neura…
▽ More
Localization is a critical technology for various applications ranging from navigation and surveillance to assisted living. Localization systems typically fuse information from sensors viewing the scene from different perspectives to estimate the target location while also employing multiple modalities for enhanced robustness and accuracy. Recently, such systems have employed end-to-end deep neural models trained on large datasets due to their superior performance and ability to handle data from diverse sensor modalities. However, such neural models are often trained on data collected from a particular set of sensor poses (i.e., locations and orientations). During real-world deployments, slight deviations from these sensor poses can result in extreme inaccuracies. To address this challenge, we introduce FlexLoc, which employs conditional neural networks to inject node perspective information to adapt the localization pipeline. Specifically, a small subset of model weights are derived from node poses at run time, enabling accurate generalization to unseen perspectives with minimal additional overhead. Our evaluations on a multimodal, multiview indoor tracking dataset showcase that FlexLoc improves the localization accuracy by almost 50% in the zero-shot case (no calibration data available) compared to the baselines. The source code of FlexLoc is available at https://github.com/nesl/FlexLoc.
△ Less
Submitted 10 June, 2024;
originally announced June 2024.
-
GDTM: An Indoor Geospatial Tracking Dataset with Distributed Multimodal Sensors
Authors:
Ho Lyun Jeong,
Ziqi Wang,
Colin Samplawski,
Jason Wu,
Shiwei Fang,
Lance M. Kaplan,
Deepak Ganesan,
Benjamin Marlin,
Mani Srivastava
Abstract:
Constantly locating moving objects, i.e., geospatial tracking, is essential for autonomous building infrastructure. Accurate and robust geospatial tracking often leverages multimodal sensor fusion algorithms, which require large datasets with time-aligned, synchronized data from various sensor types. However, such datasets are not readily available. Hence, we propose GDTM, a nine-hour dataset for…
▽ More
Constantly locating moving objects, i.e., geospatial tracking, is essential for autonomous building infrastructure. Accurate and robust geospatial tracking often leverages multimodal sensor fusion algorithms, which require large datasets with time-aligned, synchronized data from various sensor types. However, such datasets are not readily available. Hence, we propose GDTM, a nine-hour dataset for multimodal object tracking with distributed multimodal sensors and reconfigurable sensor node placements. Our dataset enables the exploration of several research problems, such as optimizing architectures for processing multimodal data, and investigating models' robustness to adverse sensing conditions and sensor placement variances. A GitHub repository containing the code, sample data, and checkpoints of this work is available at https://github.com/nesl/GDTM.
△ Less
Submitted 21 February, 2024;
originally announced February 2024.
-
Using DeepProbLog to perform Complex Event Processing on an Audio Stream
Authors:
Marc Roig Vilamala,
Tianwei Xing,
Harrison Taylor,
Luis Garcia,
Mani Srivastava,
Lance Kaplan,
Alun Preece,
Angelika Kimmig,
Federico Cerutti
Abstract:
In this paper, we present an approach to Complex Event Processing (CEP) that is based on DeepProbLog. This approach has the following objectives: (i) allowing the use of subsymbolic data as an input, (ii) retaining the flexibility and modularity on the definitions of complex event rules, (iii) allowing the system to be trained in an end-to-end manner and (iv) being robust against noisily labelled…
▽ More
In this paper, we present an approach to Complex Event Processing (CEP) that is based on DeepProbLog. This approach has the following objectives: (i) allowing the use of subsymbolic data as an input, (ii) retaining the flexibility and modularity on the definitions of complex event rules, (iii) allowing the system to be trained in an end-to-end manner and (iv) being robust against noisily labelled data. Our approach makes use of DeepProbLog to create a neuro-symbolic architecture that combines a neural network to process the subsymbolic data with a probabilistic logic layer to allow the user to define the rules for the complex events. We demonstrate that our approach is capable of detecting complex events from an audio stream. We also demonstrate that our approach is capable of training even with a dataset that has a moderate proportion of noisy data.
△ Less
Submitted 15 October, 2021;
originally announced October 2021.
-
Improved Techniques for the Surveillance of the Near Earth Space Environment with the Murchison Widefield Array
Authors:
Brendan Hennessy,
Steven Tingay,
Paul Hancock,
Robert Young,
Steven Tremblay,
Randall B. Wayth,
John Morgan,
Sammy McSweeney,
Brian Crosse,
Melanie Johnston-Hollitt,
David L. Kaplan,
Dave Pallot,
Mia Walker
Abstract:
In this paper we demonstrate improved techniques to extend coherent processing intervals for passive radar processing, with the Murchison Widefield Array. Specifically, we apply a two stage linear range and Doppler migration compensation by utilising Keystone Formatting and a recent dechir** method. These methods are used to further demonstrate the potential for the surveillance of space with th…
▽ More
In this paper we demonstrate improved techniques to extend coherent processing intervals for passive radar processing, with the Murchison Widefield Array. Specifically, we apply a two stage linear range and Doppler migration compensation by utilising Keystone Formatting and a recent dechir** method. These methods are used to further demonstrate the potential for the surveillance of space with the Murchison Widefield Array using passive radar, by detecting objects orders of magnitude smaller than previous work. This paper also demonstrates how the linear Doppler migration methods can be extended to higher order compensation to further increase potential processing intervals.
△ Less
Submitted 1 July, 2019;
originally announced July 2019.
