Lorenzo Shaikewitz

Graduate Student, MIT AeroAstro

I am a graduate student at MIT AeroAstro advised by Prof. Luca Carlone. I'm broadly interested in bridging the gap between navigation and interaction, enabling robots to work in human spaces.

Previously, I was a Caltech undergrad (Mechanical Engineering & Controls, '23) advised by Prof. Aaron Ames and Prof. Günter Niemeyer. My undergraduate projects included a compliant ankle exoskeleton (with Maegan Tucker & Prof. Ames), an in-mouth feeding robot (with Prof. Dorsa Sadigh at Stanford), and a piano playing robot for ME 134 (with Prof. Niemeyer).

email / scholar / github / youtube

Graduate Research

Certifiable Category-Level Tracking  //  June 2024
My first publication with SPARK Lab considers the problem of simultaneously estimating an object's shape and tracking its pose. We propose a certifiably optimal method built on the convex relaxation of a fixed lag smoother. We show the method largely outperforms others on simulated data, the YCBInEOAT manipulation dataset, and a custom drone vehicle tracking scenario.

Read our paper on arXiv and view our code on github.

Undergraduate Work

Interactive Piano Playing Robot  //  March 2023
For ME 134 (Robot Systems) my team built a piano playing robot. We designed a 9 degree of freedom robot with two end effectors capable of playing up to two notes at a time. The emphasis was interactivity: our robot followed the piano if someone moved it around, even pulling it back into a playable region if necessary. It could teach the user to play a song, halt playing when someone got too close, and celebrate with claps and dances when it finished.
Robotic In-Mouth Bite Transfer  //  September 2022
Over summer 2022 I worked in Prof. Dorsa Sadigh's lab at Stanford on robotic bite transfer. During the fellowship I designed and 3D printed a compact two degree-of-freedom wrist mount and used it to make robotic assistive feeding more comfortable. My focus was bringing bite-sized food directly into the mouth.

Read our paper on arXiv or IEEE.
Line-Based Mapping  //  June 2022
For ME 169, a mobile robot project class, I worked on a continuous-space mapping alogorithm using the robot's LiDAR system. Rather than work on a grid, as most SLAM algorithms do, this method stored the measured coordinates of walls as line segments in continuous space. At each time step, we converted the LiDAR point cloud into line segments using a least-squares regression. These measurement lines were combined with previously measured lines to build up a map of a space.
MechE Capstone Design  //  March 2022
I was part of a team of seven that competed in the 37th Annual ME 72 Competition for our capstone design. The competition was about building sumo robots--together we built one wheeled robot capable of R/C and autonomous sumo flight and one "strand bot" that used a planar linkage inspired by Theo Jansen's Strandbeests. I focused on the latter bot, which is shown in battle left.
Strut Ankle Exoskeleton  //  October 2020
My first research experience was working on the design of a strut-based ankle exoskeleton. This concept, similar to this design developed at MIT, was originally going to be manufactured over the summer. I used the summer of quarantine to optimize the device design, learn to make printed circuit boards, and go through basic control simulations.