Jogging Stroller Mechanics

Running while pushing a jogging stroller is common, yet the biomechanical demands of stroller running remain poorly understood. Our research program examines how stroller use changes running mechanics, loading, muscle activity, and movement strategies using a combination of laboratory biomechanics, wearable sensors, and field-based data collection.

Our published work in PLOS ONE demonstrated that stroller running produces a biomechanical tradeoff: several measures of vertical impact loading were reduced while torsional loading increased. Rather than simply increasing or decreasing running loads, pushing a stroller appears to change how those loads are distributed.

We have extended this work through conference presentations examining ground reaction forces, free moment, 3D joint kinematics, tibial acceleration, muscle activation, arm swing and trunk rotation, hands-free stroller running, stride characteristics, running volume and injury risk, and stroller running on hills. This growing body of work allows us to examine stroller running across multiple biomechanical systems rather than through a single outcome measure.

Current studies are expanding the project beyond traditional laboratory conditions. We are collecting outdoor data to determine how hills and changes in grade affect stroller-running mechanics and loading. A second laboratory study is comparing different stroller types and examining how restrictions on arm swing alter running biomechanics. Together, these studies are helping us understand how the runner, stroller, environment, and running technique interact.

This research also serves as a platform for undergraduate researchers. Students can participate in experimental design, human-subject data collection, motion analysis, wearable sensing, signal processing, programming, statistical analysis, and scientific communication. Individual student projects can address focused questions while contributing to a larger research program, creating opportunities for conference presentations, publications, and continued collaborative work.

Our long-term goal is to develop a comprehensive understanding of stroller-running biomechanics across runners, equipment, techniques, and environments while building a sustainable interdisciplinary research program that supports undergraduate research and collaboration. The project provides a growing foundation for work spanning biomechanics, engineering, rehabilitation, wearable technology, human movement science, and product design.

See our peer-reviewed publication in PLoS One 

• https://doi.org/10.1371/journal.pone.0332616 

See our conference presentations

• Running Volume and Stroller Use: Their Combined Effect on Injury Risk

• Stroller Running Slightly Attenuates the Stride Length–Speed Relationship

• How Does Running with a Jogging Stroller Affect Ground Reaction Force?

• Muscle Activation During Stroller Running

• Free Moment Increases While Running with a Stroller

• Does a Running Stroller Affect 3D Joint Kinematics?

• How Does a Running Stroller Affect Tibial Acceleration During a Run? 

• How Does Hands-Free Stroller Running Affect the Ground Reaction Force?

• Arm Swing, Trunk Rotation, and Free Moment During Running

• Stroller Running on Hills: How Terrain Affects Ground Reaction Forces

and our articles in the Washington Post and The Conversation 

• What running with a jogging stroller does to your body

• Running with a stroller: 2 biomechanics researchers on how it affects your form − and risk of injury 

Our stroller running work has also been featured in the Newsweek article

• Surprising Effect of Running With a Stroller Revealed.

Schmear: Automated OpenCap Gait Analysis Pipeline

Markerless motion capture systems such as OpenCap have dramatically lowered the barrier to collecting biomechanical data. Standard video cameras can now generate detailed estimates of human movement without the cost and infrastructure of a traditional motion capture laboratory. However, collecting the data is only part of the challenge. Turning OpenCap outputs into interpretable gait measures still requires substantial processing, technical expertise, quality control, and manual interpretation.

Schmear was developed to help bridge that gap. The pipeline was originally created in MATLAB to process OpenCap gait data, identify gait events, organize biomechanical measures, and automate analysis and reporting. We are now rebuilding the pipeline in Python with a Jupyter-based interface to improve accessibility, portability, reproducibility, and opportunities for continued development.

The developing workflow is designed to support multiple users. A streamlined clinical pathway produces concise PDF reports that summarize key gait measures, while a research pathway provides more detailed outputs for quality control, statistical analysis, and further investigation. An important part of the redevelopment is preserving the analysis developed in MATLAB while creating a more flexible framework for adding new measures and methods.

Schmear also provides a platform for undergraduate research at the intersection of biomechanics, computing, and data science. Student projects can contribute to gait-event detection, signal processing, software validation, data visualization, quality-control methods, and the development of new biomechanical measures. The modular nature of the pipeline also creates opportunities for collaboration with biomechanics researchers, clinicians, rehabilitation specialists, and others interested in making quantitative gait analysis more accessible.

Our longer-term goal is to develop Schmear into a robust research framework that helps translate increasingly accessible markerless motion capture data into useful and reproducible biomechanical information. This work was presented at the World Congress of Biomechanics:

Bridging OpenCap Outputs to Clinically Useful Gait Reports

Our research group works with undergraduate and graduate students on projects in biomechanics, gait analysis, wearable sensing, markerless motion capture, machine learning, and computational methods for studying human movement.

Student Opportunities

Most of our student research takes place during the summer in a collaborative environment that brings together students from different majors, institutions, academic interests, and stages of their education. Students from early undergraduate years through graduate study work alongside one another and with faculty, contribute to shared projects, and also develop focused questions of their own.

Students learn how to move a research project from an idea to a finished product. This includes study design, literature review, human-subject data collection methods, Vicon post-processing, data analysis, and scientific communication. Students also gain experience presenting research posters and contributing to the writing of conference abstracts, reports, and peer-reviewed papers.

One of our longer-term goals is to grow externally funded summer research opportunities that bring together students from multiple institutions and disciplines to work on connected research problems.

Collaborations

We are interested in collaborations that can help expand this work, including partnerships with researchers in biomechanics, rehabilitation, human movement science, wearable technology, and accessible measurement methods.

We are also interested in working with stroller manufacturers and product-design teams to better understand how stroller design affects running biomechanics and user experience, and to develop research questions that are useful both scientifically and in product development.

Students, faculty, and potential research or industry collaborators who are interested in this work are welcome to reach out.

Footstrike Pattern Identification

Footstrike pattern is commonly classified as rearfoot, midfoot, or forefoot and can influence running mechanics and loading. Traditional classification often relies on visual assessment or laboratory measurements, which can be subjective and difficult to apply consistently across large datasets. This project used machine-learning methods to automatically identify footstrike pattern from biomechanical data. Quantitative features extracted from running data were used to distinguish among footstrike patterns, providing a more objective and scalable alternative to manual classification. The work resulted in a peer-reviewed publication in the Journal of Biomechanics and helped establish methods in machine learning, signal processing, and automated gait analysis that continue to inform our current research.

Associated Publication:

Footstrike Pattern Recognition Using Machine Learning on Tibial Accelerometry

Treadmill Retrofit

Instrumented treadmills can provide valuable measurements such as ground reaction forces, center of pressure, and gait-event timing, but commercial systems are expensive and can be difficult to justify for smaller research programs. This project retrofitted a conventional treadmill with load cells to create a lower-cost platform for biomechanics research. A custom PCB interfaced the sensors with a Vicon data-acquisition system, and post-processing methods were developed to convert the measured signals into force and center-of-pressure estimates. The project helped expand our experience in low-cost instrumentation, sensor integration, and biomechanics data acquisition, and it contributed to our broader interest in making quantitative human-movement measurement more accessible.

Immersive Virtual Reality System

This project explored how immersive virtual reality could be integrated with human-movement research to create controlled, interactive environments for studying locomotion and user response. The system combined a virtual reality headset, motion tracking, treadmill-based movement, and custom software to create an environment in which participants could move through and interact with a simulated space while their physical motion was measured. The work involved hardware integration, programming, synchronization, and experimental testing. The project helped build experience in immersive technology, real-time motion tracking, human-computer interaction, and the integration of virtual environments with biomechanics research tools.

Virtual Shuffleboard and Air Hockey 

This project explored how low-cost sensing and interactive technology could be used to recreate physical tabletop games in a digital environment. Virtual versions of shuffleboard and air hockey were developed to study user interaction, motion tracking, and real-time system response. The work combined sensor integration, programming, graphical interfaces, and experimental testing to translate physical movement into responsive virtual gameplay. Students were involved in system design, software development, hardware integration, and evaluation of the user experience. Although outside the main biomechanics focus of our current work, these projects helped build experience in human-computer interaction, motion sensing, real-time data processing, and the development of accessible interactive systems.

Shuffleboard Mark I

Shuffleboard Mark II

Early Prototype Version (Summer 2016): Video courtesy MA Bianco

Thesis-Ready Mark I (Spring 2018)

More technical information is available from the ASB 2018 Regional conference and "A low-cost, open-source virtual air hockey table for human motion applications"

Mahoney Dissertation - https://etda.libraries.psu.edu/catalog/19584

Bianco Thesis - https://honors.libraries.psu.edu/catalog/5326mab6756

Step Counting and Evaluation

Consumer activity trackers made step counting widely accessible, but their accuracy can vary with walking and running speed, sensor placement, and type of movement. This work examined how low-cost wearable sensors and simple computational methods could be used to measure steps and characterize gait more reliably. We began by developing and validating open-source step-counting algorithms using low-cost IMUs and comparing different sensor locations and methods. The work then expanded beyond simply counting steps to identifying how a person was moving, including distinguishing walking from running and estimating additional characteristics of gait. These projects resulted in peer-reviewed work on step counting, gait classification using machine learning, optical validation methods, and joint-angle estimation from IMUs. This research also helped establish the wearable-sensing, signal-processing, and machine-learning methods that continue to inform our work in automated gait analysis and footstrike identification.

Associated Peer-Reviewed Articles:

• Methodology and validation for identifying gait type using machine learning on IMU data

• A comprehensive comparison of simple step counting techniques using wrist- and ankle-mounted accelerometer and gyroscope signals

• Design and validation of a simple automated optical step counting method for treadmill walking

• Knee Angle Estimation with Dynamic Calibration Using Inertial Measurement Units for Running

Inexpensive and Portable 2D Motion Capture

Traditional motion-capture systems provide detailed biomechanical measurements but are expensive, require dedicated laboratory space, and can be difficult to deploy outside controlled research environments. This project explored whether lower-cost optical methods could provide useful motion data with simpler and more portable equipment. We developed and evaluated an inexpensive planar motion-capture approach using commonly available cameras and open-source or low-cost computational tools. The work focused on marker tracking, joint-angle estimation, synchronization, and validation against established measurement methods. The project resulted in conference presentations and a validation study demonstrating the potential of low-cost optical motion capture for biomechanics and engineering applications. It also helped establish the accessible-measurement and computational methods that now inform our work in markerless motion capture, OpenCap, and automated gait analysis.

More technical information available from the ASB 2015 conference. The detection algorithm was adapted for the aforementioned step counting work.

Daily Fantasy Football Roster Optimization

Daily fantasy sports provide an interesting applied optimization problem: selecting the strongest possible roster while satisfying salary-cap, position, and lineup constraints. This project explored computational methods for constructing daily fantasy football lineups using player projections and mathematical optimization. We developed approaches for generating optimized rosters and examined how different projections, constraints, and lineup-building strategies affected roster selection. The work provided an accessible application of operations research, programming, data analysis, and decision-making under uncertainty. Although outside the primary biomechanics focus of our current research, this project reflects a broader interest in using computational and optimization methods to solve practical problems with large datasets and competing constraints.

Method and Validation for Optimal Lineup Creation for Daily Fantasy Football Using Machine Learning and Linear Programming

Continuous Vital Sign Monitor

Continuous monitoring of physiological signals can provide useful information about health and human performance, but many commercial systems are expensive, proprietary, or difficult to adapt for research. This project explored a lower-cost approach to collecting and monitoring vital-sign data using custom instrumentation and embedded systems. The work focused on integrating multiple sensors into a portable monitoring platform, acquiring physiological signals continuously, and developing the hardware and software needed to record and interpret those data. The project combined circuit design, sensor integration, signal processing, and data acquisition. Although this work predates our current biomechanics projects, it helped establish an interest in wearable sensing, low-cost measurement systems, and the use of engineering tools to make human-subject data collection more accessible.

Balance Board for Quiet Stance Testing 

Quiet-stance testing is commonly used to evaluate balance and postural control, but laboratory-grade force platforms can be expensive and limit access to this type of measurement. This project explored whether a lower-cost balance board could provide useful center-of-pressure data for biomechanics and human-movement research. The work focused on developing and validating a portable measurement system capable of capturing postural sway during quiet standing. Students contributed to hardware development, data acquisition, signal processing, and comparison of the system against established laboratory methods. This project helped extend our work in low-cost instrumentation and accessible biomechanics by showing how relatively simple hardware and computational tools can be used to study human balance and postural control.

Associated Publications and Presentations: 

ASB 2016 conference.

Instrumented Bicycle

Cycling provides a useful platform for studying human performance, movement, and mechanical efficiency, but many of the measurements used in laboratory testing require specialized equipment. This project focused on developing an instrumented bicycle capable of collecting mechanical and physiological data during cycling. The system integrated sensors and data-acquisition hardware to measure rider and bicycle performance during controlled testing. The work combined mechanical design, instrumentation, signal processing, and human-subject data collection. This project contributed to our broader interest in developing practical measurement systems for human movement and performance and helped build experience in sensor integration, experimental design, and applied biomechanics.

Research Media and Features

Selected media coverage and institutional features highlighting the impact of our biomechanics research, undergraduate research mentorship, and efforts to make human-movement analysis more accessible. See News for broader updates on engineering education and the John R. Post School of Engineering.