Email: zhanteng.xie (at) temple (dot) edu
Centre for Transformative Garment Production, HK : Jun. 2025 - Oct. 2025
Postdoctoral Fellow
Temple University, US : Aug. 2019 - Aug. 2024
Research/Teaching Assistant
Southern University of Science and Technology, CN : Jun. 2018 - Jul. 2019
Research Assistant
Temple University, US : Aug. 2019 - Aug. 2024
Doctor of Philosophy in Mechanical Engineering (GPA=4.0)
Dissertation title: Robot Navigation in Crowded Dynamic Scenes
Harbin Institute of Technology, CN : Sept. 2015 - Jan. 2018
Master of Engineering in Information and Communication Engineering
Zhengzhou University, CN : Sept. 2011 - Jul. 2015
Bachelor of Engineering in Electronic Information Engineering
- Supervised by Prof. Philip Dames Sept. 2019 - Present
How to simulate the real world: Designed and implemented a human-robot interaction 3D Gazebo simulator for training and testing neural network-based navigation policies.
How to process perceived sensor data and obtain scene understanding: Proposed the first 2D lidar semantic segmentation dataset (Semantic2D), a semi-automatic semantic labeling framework, and a hardware-friendly stochastic semantic segmentation algorithm (S3-Net) to fill the gap between 3D lidar semantic segmentation and 2D lidar semantic segmentation enhance the semantic scene understanding for mobile robots in different indoor robotics applications.
How to predict complex and stochastic environment states: Proposed and implemented an occupancy grid map (OGM) prediction dataset and a family of software-optimized stochastic occupancy grid map predictors (SCOPE family) by leveraging the motion information (robot, dynamic objects, and static scene) and variational autoencoders, which is tested on simulated and real-world datasets collected by three different robots, as well as navigation experiments with real-world robots.
How to efficiently model environmental states and represent them in an efficient and concise abstraction: Proposed and designed novel environment representations that are specialized for crowded dynamic spaces and allow good generalization to new environments by abstractly modeling environmental states, which are evaluated by an imitation learning-based control policy for robot navigation.
How to design crowd-aware control policies that balance safety and speed and improve generalizability: Proposed and implemented a deep reinforcement learning-based control policy (DRL-VO) by designing a novel velocity obstacle-based reward function, which is tested in both simulations and real-world robotics experiments.
How to design an uncertainty-aware navigation framework that improves the safe navigation performance of existing control policies: Proposed and implemented a hardware-friendly predictive uncertainty-aware navigation framework (SCOPE-PU series) by incorporating prediction and its uncertainty information from the SCOPE family into current existing model-based and learning-based control policies, which is tested in both simulation and real-world robotics experiments.
How to design socially-aware control policies that learn to work with people and improve human comfort: In progress.
Github repositories:
[3D human-robot interaction Gazebo simulator], [Semantic2D], [Stochastic occupancy grid map predictor], [Crowd-aware control policy], [Uncertainty-aware control framework] .
-------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. Philip Dames Jun. 2020 - Mar. 2021
How to design and create datasets for computer vision: Collected and labeled a water bottle detection dataset for tiny object detection and tracking.
How to fine-tune a general-purpose computer vision detector for a specific robotics application: Fine-tuned a YoloV3 object detector using the custom dataset to detect different brands of water bottles, which is tested in real-world robotics experiments.
How to characterize sensor models for probabilistic filters: Proposed and implemented an experimental characterization method for range-bearing sensor models (observation, detection, and clutter) in probability hypothesis density (PHD) filters, which is tested in real-world experiments.
Github repositories:
[Bottle detector], [Sensor models].
-------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. Philip Dames Sept. 2020 - Oct. 2021
How to simulate industrial robots and their application scenarios: Designed and implemented a 3D industrial ground vehicle material handling scene simulator, simulating an industrial ground vehicle (FRED2500) navigating among a magnetic tape ``roadway".
How to communicate and control industrial robots: Designed and implemented a management system by adding a high-level ROS embedded controller with additional camera sensors to control and cooperate with the low-level drive controller (Roboteq), which is tested in real-world experiments.
Github repositories:
Doctoral Dissertation:
Zhanteng Xie. “Robot Navigation in Crowded Dynamic Scenes.” Temple University. 2024.
Peer-Reviewed Journal Publications:
Pujie Xin, Zhanteng Xie, and Philip Dames. “Towards Predicting Collective Performance in Multi-Robot Teams.” IEEE Transactions on Robotics (T-RO), 2025. [arXiv]
Zhanteng Xie and Philip Dames. “SCOPE: Stochastic Cartographic Occupancy Prediction Engine for Uncertainty-Aware Dynamic Navigation.” IEEE Transactions on Robotics (T-RO), 41, 4139 - 4158, 2025. [arXiv], [video], [Github(predictor)], [Github(navigation)]
Zhanteng Xie and Philip Dames. “DRL-VO: Learning to Navigate Through Crowded Dynamic Scenes Using Velocity Obstacles.” IEEE Transactions on Robotics (T-RO), 39(4), 2700-2719, 2023. [arXiv], [video], [Github]
Xuesu Xiao, Zifan Xu, Zizhao Wang, Yunlong Song, Garrett Warnell, Peter Stone, Tingnan Zhang, Shravan Ravi, Gary Wang, Haresh Karnan, Joydeep Biswas, Nicholas Mohammad, Lauren Bramblett, Rahul Peddi, Nicola Bezzo, Zhanteng Xie, and Philip Dames. "Autonomous Ground Navigation in Highly Constrained Spaces: Lessons Learned From the Benchmark Autonomous Robot Navigation Challenge at ICRA 2022 [Competitions]." IEEE Robotics & Automation Magazine (RAM), 29(4), 148-156, 2022. [Github]
Jun Chen, Zhanteng Xie, and Philip Dames. “The Semantic PHD Filter for Multi-class Target Tracking: From Theory to Practice.” Robotics and Autonomous Systems (RAS), 149, 2022. [Github]
Peer-Reviewed Conference Publications:
Lindsay Ouellette, Julia Zortea, Alexandra Capalbo, Raz Shermister, Charles Furtado, Donald Hantula, Philip Dames, and Zhanteng Xie. “Too close for comfort? Navigating personal space in human-robot encounters.” Northeast Business & Economics Association (NBEA) Conference. 2024.
Zhanteng Xie and Philip Dames. “Stochastic Occupancy Grid Map Prediction in Dynamic Scenes.” Conference on Robot Learning (CoRL), 2023. [arXiv], [Github]
Zhanteng Xie, Pujie Xin, and Philip Dames. “Towards Safe Navigation Through Crowded Dynamic Environments.” IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021.
Publications Under Review:
Zhanteng Xie and Philip Dames. “Semantic2D: A Semantic Dataset for 2D Lidar Semantic Segmentation.” arXiv preprint arXiv:2409.09899, 2024. [arXiv], [video], [Dataset], [Github]
Lindsay Ouellette, Julia Zortea, Alexandra Capalbo, Raz Shermister, Philip Dames, Zhanteng Xie, Charles Furtado, and Donald Hantula. “Too close for comfort? Navigating personal space in human-robot encounters.” 2024.
Lindsay Ouellette, Drew Kronstadt, Sunvy Yalamarthy, Philip Dames, Zhanteng Xie, Slobodan Vucetic, and Donald Hantula. “Human vs. Robot Approaches: The Influence of Negative Attitudes on Comfort Levels.” 2024.
1st place in the 2025 ICRA Arena Challenge (social navigation competition) at ICRA 2025, 2025
3rd place in the Benchmark Autonomous Robot Navigation (BARN) Challenge at ICRA 2022 (1st place in the Simulation Portion), 2022
3rd place in the 7th Zhengzhou University ACM Programming Design Contest, 2014
2nd place in the National Undergraduate Electronic Design Contest, 2013
1st place in Henan division of the National Undergraduate Electronic Design Contest, 2013
1st class Postgraduate Scholarship, 2016
2nd class Postgraduate Scholarship, 2015
3rd class Excellent Student Scholarship, 2014
3rd class Excellent Student Scholarship, 2012
National Computer Rank Examination Certificate (C++), Level 2, 2012
IEEE Transactions on Robotics (T-RO), 2023, 2024, 2025.
Robotics and Autonomous Systems (RAS), 2024, 2025.
IEEE Robotics and Automation Letters (RA-L), 2021, 2023, 2024, 2025.
IEEE International Conference on Robotics and Automation (ICRA), 2021, 2023, 2024, 2025.
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021, 2022, 2023, 2024, 2025.
Robotic development
Languages: Python, C++, and MATLAB;
ROS frameworks: ROS1/2, PedSim, Flatland, Simulink, and Gazebo;
Learning frameworks: Pytorch, PyTorch Geometric, Stable-Baselines3, and Tensorflow;
Robots: Turtlebot3, Turtlebot2, Jackal, and ASI Fred AGV;
Embedded computing devices: Raspberry Pi 3B/4, Nvidia Jetson TX2, and Nvidia Jetson AGX Xavier;
Sensors: Hokuyo UTM-30LX lidar, ZED Stereo camera, Intel RealSense tracking camera T265, and depth camera D435.
Machine learning development
Languages: Python and C++;
Deep learning Frameworks: Pytorch, PyTorch Geometric, Tensorflow, Caffe, NCNN, and NNIE;
Deep reinforcement learning Frameworks: Stable-Baselines3.
MATLAB development
M language, S function, and Simulink platform.
FPGA development
Languages: Verilog, Testbench in Verilog;
Software: Quartus II, Xilinx ISE, Vivado, and Modelsim
FPGAs: Altera series and Xilinx series.
Embedded development
Languages: C and C++;
Software: Keil, Linux, and OpenCV;
Embedded microcontrollers: C51-SCM, STC12-SCM, STM32-SCM, and ARM series.
Electronic circuit design development
Software: Multisim, Protel 99, and Altium Designer;
Hardware: soldering PCB circuit boards;
Electronic Components: capacitors, resistors, and different types of driver ICs.
- Supervised by Prof. Hao YU Jun. 2018 - Jul. 2019
Goal - to realize the video image real-time processing in embedded platforms without GPUs, including video detection and video analysis;
Tasks -
1. To realize the real-time object detection in video using SSD/YOLO network combined with MobileNet in mobile platforms, including training full-precision network using Caffe architecture, writing related inference code using NCNN architecture and NNIE architecture, and getting 8bits quantized network using DoReFa algorithm.
2. To realize the real-time video analysis using RNN combined with tensor compression in mobile platforms.
Outcome – realized the video object detection demo on RK3399, HiKey970, and Hi3559A platforms, achieving 10 FPS processing speed on HiKey970 (full-precision network) and 18 FPS processing speed on Hi3559A (8bits quantized network).
---------------------------------------------------------------------------------------------------------------------------------------
Apr. 2016-Jun. 2017
Goal - to devise and realize the WCDMA communication system for LEO Satellites;
Tasks -
1. To realize the QPSK system via M-Language of MATLAB, and verify the BER performance of the QPSK system under the AWGN channel, Rayleigh channel, and Rice channel;
2. To set up the downlink of the WCDMA communication simulation system for LEO satellites based on the Simulink platform, which includes the realization of physical layer algorithms, and verify the system’s BER performance under the environment of low SNR and large Doppler shift;
3. To do hardware implementation of the downlink of the WCDMA communication system for LEO satellites via Verilog language based on Xilinx FPGA platforms, which includes spreading spectrum, scrambling, shaping filter, up/down-conversion, code synchronization, carrier synchronization, and AWGN channel simulator, and complete the measurement of system’s BER performance;
4. To participate in the uplink design of the WCDMA communication system for LEO satellites, including parameter designs such as physical channel rate, bandwidth and IF frequency, and the physical layer architecture design;
5. To research the CDMA uplink power control algorithms for LEO satellites, and try to use the Game theory to solve the power control problem;
Outcome - established the simulation demo system and the hardware demo system for the communication system of LEO satellites.
---------------------------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. Fei WANG Oct. 2015-Nov. 2015
Goal - to realize Faddeev’s algorithm which is use for matrix computations by using systolic arrays based on FPGA;
Tasks -
1. To realize 3x3 matrix floating point multiplication by using basic systolic arrays in Verilog language and do related simulation verifications including the conversion of data format between floating point numbers and real numbers by using testbench and Modelsim software;
2. To implement ordinary mode systolic arrays and dual mode systolic arrays for Faddeev’s algorithm via Verilog language based on Xilinx FPGA platforms and compare the resource consumption and performance between these two different systolic architectures.
Outcome – implemented 4x4 matrix floating point computations by using ordinary mode systolic arrays and dual mode systolic arrays separately.
---------------------------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. JianWei JI Oct. 2014-Mar. 2015
Goal - to generate the sinusoidal wave, square wave, and triangular wave, and set tunable square wave frequency in the range of 0-1MHz and the tunable voltage of 0-5V;
Tasks -
1. To apply C Language to edit the bottom-driver code (namely 12864 LCD driver, AD9834 driver, matrix keyboard driver, and code switch driver), manipulation function code, and multi-layer interactive menu code by using STC12-SCM;
2. To design the digital signal source panel, including its size, silk screen, and the layout of LCDs, buttons, and switches;
3. To join in circuit board welding, hardware debugging, and signal source components assembly for more than 100 sets of digital signal source generators;
Outcome - designed more than 100 sets of digital signal source generators for certain district areas.
---------------------------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. JiaYou SONG Apr. 2014-Jul. 2014
Goal - to sense weather and control the window automatically; to conduct real-time monitoring of environmental quality; to realize the function of remote control;
Tasks -
1. To complete the circuit principle design and place & route of PCB circuit board via Altium Designer software;
2. To prepare circuit boards and conduct hardware debugging;
3. To apply C Language to edit the bottom-driver code (namely TFT colorized screen driver, touch screen driver, motor driver, and various sensor drivers), GSM module communication code, and multi-layer interactive menu code by using STM32-SCM;
Outcome - established the demo system for smart windows.
---------------------------------------------------------------------------------------------------------------------------------------
- Supervised by Prof. JiCai DENG Dec. 2013-Mar. 2014
Goal - to store and update control orders and display executive results; to send instructions to LED dot-matrix screen through the serial interface;
Tasks –
1. To design and realize the data transmission protocol between the upper monitor and handle control terminal;
2. To apply C Language to edit the bottom-driver code (namely 12864 LCD driver, button driver, serial port driver, and Flash driver by SPI bus), control function code, and multi-layer interactive menu code by using STC12-SCM;
Outcome - published paper Design of The Control Handle for LED Display Screen Based on SCM (2nd Author), Journal of Henan Science and Technology (ISSN1003-5168), April 2014.