Core Challenges of Social Robot Navigation: A Survey

Mavrogiannis, Christoforos; Baldini, Francesca; Wang, Allan; Zhao, Dapeng; Trautman, Pete; Steinfeld, Aaron; Oh, Jean

doi:10.48550/arxiv.2103.05668

Cited by 19 publications

(20 citation statements)

References 196 publications

(281 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Under this condition, pedestrians act as dynamic obstacles that do not react to the robot, a situation which could arise in cases where robots are of shorter size and could thus be easily missed by navigating pedestrians. In the Online one, the robot navigates among a crowd 2 moving by running ORCA [16], a policy that is frequently used as a simulation engine for benchmarking in the social navigation literature [53,8,54].…”

Section: Methodsmentioning

confidence: 99%

Group-based Motion Prediction for Navigation in Crowded Environments

Wang¹,

Mavrogiannis²,

Steinfeld³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

We focus on the problem of planning the motion of a robot in a dynamic multiagent environment such as a pedestrian scene. Enabling the robot to navigate safely and in a socially compliant fashion in such scenes requires a representation that accounts for the unfolding multiagent dynamics. Existing approaches to this problem tend to employ microscopic models of motion prediction that reason about the individual behavior of other agents. While such models may achieve high tracking accuracy in trajectory prediction benchmarks, they often lack an understanding of the group structures unfolding in crowded scenes. Inspired by the Gestalt theory from psychology, we build a Model Predictive Control framework (G-MPC) that leverages group-based prediction for robot motion planning. We conduct an extensive simulation study involving a series of challenging navigation tasks in scenes extracted from two real-world pedestrian datasets. We illustrate that G-MPC enables a robot to achieve statistically significantly higher safety and lower number of group intrusions than a series of baselines featuring individual pedestrian motion prediction models. Finally, we show that G-MPC can handle noisy lidar-scan estimates without significant performance losses.

show abstract

Section: Methodsmentioning

confidence: 99%

Group-based Motion Prediction for Navigation in Crowded Environments

Wang¹,

Mavrogiannis²,

Steinfeld³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…3) Lack of diverse pedestrian models: All simulators except SocNavBench suffer from homogeneity as they deploy one fixed mathematical model for all pedestrians in the scene. Since the behavior of humans differ depending upon their age, demographics, gender and culture [54], [55], [56], a single pedestrian model applied to all human models is not a realistic representation of human population, especially in the public places like hospitals, airports, hotels, etc. SocNavBench is an exception as it is rooted in the real pedestrian data for simulating pedestrian behavior.…”

Section: Challengesmentioning

confidence: 99%

Simulators for Mobile Social Robots:State-of-the-Art and Challenges

Kaur¹,

Liu²,

Shi³

2022

Preprint

View full text Add to dashboard Cite

The future robots are expected to work in a shared physical space with humans [1], however, the presence of humans leads to a dynamic environment that is challenging for mobile robots to navigate. The path planning algorithms designed to navigate a collision free path in complex human environments are often tested in real environments due to the lack of simulation frameworks. This paper identifies key requirements for an ideal simulator for this task, evaluates existing simulation frameworks and most importantly, it identifies the challenges and limitations of the existing simulation techniques.First and foremost, we recognize that the simulators needed for the purpose of testing mobile robots designed for human environments are unique as they must model realistic pedestrian behavior in addition to the modelling of mobile robots. Our study finds that Pedsim ros [2] and a more recent SocNavBench framework [3] are the only two 3D simulation frameworks that meet most of the key requirements defined in our paper. In summary, we identify the need for developing more simulators that offer an ability to create realistic 3D pedestrian rich virtual environments along with the flexibility of designing complex robots and their sensor models from scratch.

show abstract

“…Social robot navigation as a research field has been drawing attention from scientists and engineers for decades. A thorough survey on this topic can be found in [16]. Understanding how human pedestrians interact with each other in crowds is core to this problem, which this work can be helpful of.…”

Section: Applicationsmentioning

confidence: 99%

Predicting Human Trajectories by Learning and Matching Patterns

Zhao

2021

Preprint

Self Cite

View full text Add to dashboard Cite

As more and more robots are envisioned to cooperate with humans sharing the same space, it is desired for robots to be able to predict others' trajectories to navigate in a safe and self-explanatory way.We propose a Convolutional Neural Network-based approach to learn, detect, and extract patterns in sequential trajectory data, known here as Social Pattern Extraction Convolution (Social-PEC).First, my supervisor, Dr. Jean Oh. She believed in my potentials when I had little to no experience in this research area, and offered me the opportunity to work with her. In the past two years, she almost gave the perfect amount of flexibility and guidance for me to explore and to focus. Beyond my research work, she also genuinely cares about my personal well-being and future career goals. She always asks us to have enough rest and take breaks and vacations, she encouraged me to take internship and pursue career directions that I am passionate about even when they are not aligned with the lab's interests, she sacrificed her personal time to meet with me daily in the beginning of the quarantine to motivate and encourage me when I needed it the most... I will never forget all of Jean's kindness and support to me. Second, Dr. Red Whittaker for offering me the full-time position where I started my journey in CMU; Dr. John M. Dolan, Dr. Kris Kitani and Dr. Chieko Asakawa for trusting me and hiring me when my positions keep being terminated for various uncontrollable factors. Third, Dr. Ji Zhang for helping me with robot building with his expertise; my lab mates Sam Shum, Advaith Sethuraman, Felix Labelle, Brendon Bolt, Xinjie Yao for their assistance and inspirations to my research.I also would like to thank my parents for all of their sacrifice, investment, expectation, support and love for me; my dearest roommates/landlords/best friends in Pittsburgh, David and Allison Lambacher, for taking care of me like a family whenever I am in need; friends from my small group and my church North Way Christian Community, for all the good time and spiritual support and accountability that I needed.Lastly, I have to give all credits to God, who created me, forgave all my sins, and loves me unconditionally.

show abstract

Core Challenges of Social Robot Navigation: A Survey

Cited by 19 publications

References 196 publications

Group-based Motion Prediction for Navigation in Crowded Environments

Group-based Motion Prediction for Navigation in Crowded Environments

Simulators for Mobile Social Robots:State-of-the-Art and Challenges

Predicting Human Trajectories by Learning and Matching Patterns

Contact Info

Product

Resources

About