
#Postdoc at @UofMaryland, focusing on #MachineLearning and #robotics. Grad of @UVI_edu, @Norfolkstate, @USouthFlorida. Pilot. #VINice 🇻🇮
Joined September 2016
- Tweets810
- Following484
- Followers221
- Likes1.1K
Excited to share our paper on the SET Perceptual Factors Framework was accepted to #PTAS! 🚗🤖
SET helps autonomous systems identify and model how weather, glare, sensor limits, and other factors impact perception-making safety risks more transparent and building public trust.
Thank you Dr. Morehouse and #FEF for this opportunity, especially Taylor Rodman, Ph.D. (@itsDrRodman) for the invitation! See everyone again next year! 🤞🏾
#McKnightDoctoralFellowship #McKnightFellow #McKnightGrad #FloridaEducationFund #USF #USFgrad #GoBulls
“To #localize? Or not to localize?”
I will address this question at the 2024 #IEEE #SSRR conference today in Session 2! 🤖
This starts a line of research challenging our traditional “sense => localize => plan => act” paradigm & proposes that we localize only when necessary. 🤷🏾♂️
Thank you to:
- My collaborators at @umdcs: Kasra Torshizi and @ptokekar!
- My funders: @UofMaryland #Presidents #Postdoc Fellowship, @UMD_postdocs, & #RISEUPP (riseupp.org/).
- #YiannisAloimonos (from @prgumd) for the Shakespeare-inspired line!
Troi Williams retweeted
🚀 @troidre will present at the landmark #ICRA40 @ieee_ras_icra 🇳🇱
📅 Sep 24 1st interactive session
🤖 Curious about robot localization? We use POMDPs to decide when robots should localize, especially when the traditional paradigm is slow!
w/ Kasra Torshizi & @ptokekar
Troi Williams retweeted
🚀 @troidre will present at the landmark #ICRA40 @ieee_ras_icra 🇳🇱
📅 Sep 24 1st interactive session
🤖 Curious about robot localization? We use POMDPs to decide when robots should localize, especially when the traditional paradigm is slow!
w/ Kasra Torshizi & @ptokekar
🤖 Curious about novel work in robot localization and decision-making?
Our research challenges the traditional paradigm of robot localization. In short, we used #POMDP to decide when robots should localize, especially when the traditional paradigm is not advantageous!
🚀 Thrilled to announce: I'm presenting at the landmark 40th IEEE #ICRA in Rotterdam, Netherlands! 🇳🇱
📅 Mark your calendars for Sep 24th during the first interactive session!
Troi Williams retweeted
PhDs can be great fun, but don’t need to be a Marathon.
How to complete your PhD in 3 years:
Troi Williams retweeted
Proud to host these amazing Science Terp alums today for an insightful panel discussion! @Namzo098, @kyeishalaurence, Willie May, Asha Willis and Sherita Hill Golden, we are so proud of all you have accomplished and can't wait to see what's next for you!
Won best poster at @MultiRobotSys (#IEEE #MRS2023) for the work titled “When to Localize?”! Thanks to the amazing people who stopped by the poster to talk!
Thanks also to my mentor @ptokekar, my host institution @umiacs at #UMD, and @UMD_postdocs and @CRAtweets for funding!
Troi Williams retweeted
A very cool profile of my student, Harnaik Dhami. Very proud of the work Naik’s been doing.
Naik is a rockstar “full-stack” roboticist working on planning and perception for aerial robots. He’ll be graduating next year, and is on the industry job market!
This quoted post is unavailable.
Troi Williams retweeted
Excited to be at #IROS2023 with my team! We’ll be presenting four papers. Details in the thread below!
How can we find collision- and occlusion-free #robot (sensor) states that minimize a task's error (like #localization) AND consider the environment's effect on the sensor? 🤔
Check out our latest work, #DyFOS.
Paper: arxiv.org/abs/2211.16721
Join us at #MRS2023 to learn more!
Collaborators:
Po-Lun Chen, Sparsh Bhogavilli, Vaibhav Sanjay, and @ptokekar
Thank you #NSF #CIFellows (funding), #RaasLab, and @umiacs.

















