Siddharth Paleti

Undergraduate Student

svp6245@psu.edu

Research

Using public datasets of handwriting and driving containing both writer/driver and task labels, I will perform Principal Component Analysis (PCA) on the Probabilistic Movement Primitives (ProMPs) weights for each user and task combination. I am identifying the task-invariant and user-invariant by applying an ANOVA to different cross-sections of the data (e.g., single writer across tasks and single task across writers). At the same time, my graduate student mentor will use a machine learning based method for dimensionality reduction and will use the same analysis to identify task-specific and user-specific latents. This will allow us to compare these two approaches to identifying user-specific latent parameters. Ultimately, we will use the user-specific latent parameter to generate user-specific trajectories from generic task representations (represented as ProMPs) and compare them to the actual trajectories from the datasets.

Education

B.S. Mechanical Engineering, Penn State Universty Park, 05/2027

External Projects

Constructing Open Source Micro Spot Robot
S. Paleti
2024.