BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Robotics Institute Carnegie Mellon University - ECPv6.15.12.1//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Robotics Institute Carnegie Mellon University
X-ORIGINAL-URL:https://www.ri.cmu.edu
X-WR-CALDESC:Events for Robotics Institute Carnegie Mellon University
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20250309T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20251102T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20260308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20261101T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20270314T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20271107T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260310T150000
DTEND;TZID=America/New_York:20260310T163000
DTSTAMP:20260922T085510
CREATED:20260302T172046Z
LAST-MODIFIED:20260302T172046Z
UID:150508-1773154800-1773160200@www.ri.cmu.edu
SUMMARY:Accessible Dexterous Manipulation with Soft Hands: Designs\, Methods\, Models\, and the DexKit Platform
DESCRIPTION:Abstract: \nRobot dexterity remains an open challenge in robotics that has the potential to transform manufacturing\, healthcare\, and daily life. Robots that safely and robustly interact with unstructured environments must combine compliant hardware with models and planners that tolerate uncertainty. Additionally\, if robust robot dexterity is to be realized outside of research labs\, it must be accessible to a broad audience\, with low-cost hardware and open-source software. \nThis thesis advances dexterous manipulation with soft\, tendon-driven hands by integrating new fabrication and control methods\, data-driven models of manipulation capabilities\, a fast algorithm for robust grasp synthesis\, and a first-of-its-kind accessible experimental platform. \nI introduce fully soft foam hands actuated by tendons routed on textile skins. I detail a simple molding-and-casting pipeline\, validate a soft-body simulation framework\, compare inverse-kinematics control strategies\, and optimize nontrivial tendon routings. I further report a user study on human-designed routings\, demonstrations of power/precision grasps and in-hand manipulation\, sub-millimeter repeatability\, and year-long durability\, alongside an analysis of limitations (e.g.\, routing through foam\, sensing\, and sim-to-real gaps). \nTo tackle the challenge of planning with soft hands\, I develop data-driven models of manipulation capabilities that capture the inherent uncertainty and redundancy of soft hands. Additionally I demonstrate a fast\, anytime method to compute globally optimal Independent Contact Regions (ICRs) by iteratively building an incremental Delaunay triangulation over grasp configuration space. I show that ICRs guide simple policies that remain robust to real-world uncertainties in object size\, pose\, and geometry. \nFinally\, to promote accessibility\, I contribute the DexKit system\, a low-cost\, anthropomorphic system (12 actuated DoF hand on a 4-DoF gantry) that can be built for under $2000. \n\nTheses Committee Members: \nNancy Pollard (chair) \nMatthew Mason \nChristopher Atkeson \nJames Bern (Williams College) \n\nA draft of the thesis document is available here.
URL:https://www.ri.cmu.edu/event/accessible-dexterous-manipulation-with-soft-hands-designs-methods-models-and-the-dexkit-platform-2/
LOCATION:3305 Newell-Simon Hall
CATEGORIES:PhD Thesis Defense,Student Talks
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260312T140000
DTEND;TZID=America/New_York:20260312T153000
DTSTAMP:20260922T085510
CREATED:20260309T220401Z
LAST-MODIFIED:20260309T220401Z
UID:150571-1773324000-1773329400@www.ri.cmu.edu
SUMMARY:Unlock Robust Spatial Perception: Towards Resilient State Estimation and Mapping for Long-Term Autonomy
DESCRIPTION:Abstract\nAutonomous robots should maintain resilient spatial perception despite sensor degradation. Humans preserve spatial awareness when moving between well-lit and dark spaces or when vision is partially occluded. Neuroscience studies suggest this robustness relies in part on a proprioception-first sensory hierarchy: the vestibular system provides continuous inertial reference signals\, while vision supplies corrective updates that refine spatial estimates (Velez-Fort et al.\, Cell\, 2025). In most robotic perception systems\, this hierarchy is reversed. Exteroceptive sensors (cameras\, LiDAR) are treated as primary\, while the IMU plays a supporting role via preintegration and gap filling. As a result\, reliability often depends on environmental conditions favorable to exteroceptive sensing. In dense smoke\, darkness\, or geometrically repetitive environments\, multiple exteroceptive modalities can degrade simultaneously\, and adding more exteroceptive sensors may not resolve these correlated failure modes. \nThis thesis asks whether we can build state estimation systems that (i) architecture design for resilient odometry (ii) estimate their own uncertainty for different modalities\, (iii) adapt fusion strategies based on current sensor reliability. \nWe first address architectural structure for continuity under exteroceptive degradation. We propose an IMU-centric factor graph (Super Odometry\, IROS 2021) in which inertial sensing provides a continuous backbone and exteroceptive sensors contribute constraints only when their measurements are sufficiently informative. This design was evaluated on aerial\, wheeled\, and legged platforms during the DARPA Subterranean Challenge. However\, enabling or disabling an entire modality is a coarse response: in a long corridor\, for example\, LiDAR scan matching can be well constrained perpendicular to the walls but poorly constrained along the corridor axis. To address this\, we develop per-axis degeneracy prediction (SuperLoc\, ICRA 2025)\, which decomposes each scan’s information matrix into six directions corresponding to the pose degrees of freedom\, enabling the optimizer to weight each measurement direction according to its estimated observability. \nNext\, we consider conditions where image measurements themselves become unreliable\, such as dense smoke\, low light\, or overexposure. We estimate per-feature covariance from the photometric Hessian across RGB and thermal modalities (MSO\, arXiv 2025)\, allowing the factor graph to weight individual features based on their estimated reliability rather than relying on a fixed noise model. When exteroceptive sensing is severely degraded or unavailable\, the IMU may be the only remaining signal. Standard preintegration models\, which assume slowly varying biases\, can accumulate substantial drift. We address this with a learned inertial odometry model (TartanIMU\, CVPR 2025) pretrained on data from diverse robot platforms\, and we enable online adaptation to previously unseen platforms through self-supervised learning. \nFinally\, we integrate these components into a unified system and evaluate it at scale. We present a hierarchical adaptation framework (Super Odometry 2.0\, Science Robotics) with four levels: per-feature uncertainty weighting\, per-axis observability adjustment\, sensor-level reconfiguration\, and learned inertial odometry as a fallback. We evaluate the system over approximately 200 km and 800 hours of operation across aerial\, wheeled\, legged\, and handheld platforms. To support standardized evaluation of SLAM robustness\, we also introduce SubT-MRS (CVPR 2024)\, a multi-robot\, multi-degradation benchmark with associated robustness metrics\, which served as the evaluation framework for the ICCV 2023 SLAM Challenge. All these works are presented on superodometry.com \n\nThesis Committee\nSebastian Scherer (Chair)\nMichael Kaess\nShubham Tulsiani\nJakob Engel (Meta) \nPlease find the PhD thesis defense here.
URL:https://www.ri.cmu.edu/event/unlock-robust-spatial-perception-towards-resilient-state-estimation-and-mapping-for-long-term-autonomy/
LOCATION:GHC 4405
CATEGORIES:PhD Thesis Defense,Student Talks
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260320T093000
DTEND;TZID=America/New_York:20260320T110000
DTSTAMP:20260922T085510
CREATED:20260316T161507Z
LAST-MODIFIED:20260316T161507Z
UID:150623-1773999000-1774004400@www.ri.cmu.edu
SUMMARY:Bring a Hand to The Sky: Towards Universal Aerial Manipulation
DESCRIPTION:Abstract:\n\nUncrewed Aerial Vehicles (UAVs) have attracted significant attention in applications such as inspection and maintenance. Many of these tasks require aerial robots to physically interact with the environment\, motivating the emerging field of aerial manipulation. However\, most existing approaches focus on a single task with specialized hardware and control strategies\, limiting their ability to generalize across tasks and environments.\n\nThis thesis investigates how to achieve universal aerial manipulation: enabling a general aerial manipulation platform to perform diverse manipulation tasks across different environments using a unified hardware and algorithmic pipeline.\n\nWe first study a fundamental capability for aerial manipulation: stable physical interaction with the environment. We integrate tactile sensing on a fully actuated hexarotor and develop a flight controller that leverages contact feedback for stabilization. Building on this capability\, we address a more general aerial interaction problem that requires simultaneous tracking of motion and contact force trajectories through a pipeline combining a contact-aware trajectory planner and a hybrid motion-force controller.\n\nWe then move beyond individual interaction tasks and develop a unified aerial manipulation framework for versatile task execution. The framework adopts an end-effector-centric interface that connects high-level policy and low-level control\, enabling intuitive aerial teleoperation and learning from demonstrations. To further scale policy learning from human demonstrations\, we propose Embodiment Aware Diffusion Policy (EADP) for embodiment-aware deployment of embodiment-agnostic manipulation policies. By integrating a diffusion policy with an embodiment-specific controller during inference\, this method guides trajectory generation toward dynamically feasible behaviors that respect the physical constraints of the aerial platform.\n\nFinally\, we introduce AM Bench\, an open source simulation suite and benchmark for aerial manipulation. It provides a modular simulation environment supporting multiple aerial manipulator embodiments\, diverse manipulation tasks\, configurable disturbances\, and standard control and imitation learning implementations\, enabling reproducible research and systematic comparison of aerial manipulation algorithms.\n\nTogether\, the unified aerial manipulator platform\, robust and precise control\, and embodiment-aware policy learning presented in this thesis progressively expand aerial robots from performing isolated interaction tasks to executing versatile manipulation behaviors across diverse scenarios. These advances move aerial robotics toward the long-term vision of universal aerial manipulation\, transforming aerial robots from passive flying sensors into active flying hands in the sky\, capable of interacting with the world with intelligence and precision.\n\n \nThesis Committee Members:\nGuanya Shi\, Carnegie Mellon University\, Co-Chair\nSebastian Scherer\, Carnegie Mellon University\, Co-Chair\nOliver Kroemer\, Carnegie Mellon University\nGiuseppe Loianno\, New York University\nJunyi Geng\, Penn State University\n\n \nPlease find the PhD thesis doc here:\nhttps://drive.google.com/drive/folders/1hkLM1UL1TuJFU3EphLgUG6vVE_WvuZfb?usp=sharing
URL:https://www.ri.cmu.edu/event/bring-a-hand-to-the-sky-towards-universal-aerial-manipulation/
LOCATION:Gates Hillman Center 4405
CATEGORIES:PhD Thesis Defense,Student Talks
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260325T110000
DTEND;TZID=America/New_York:20260325T123000
DTSTAMP:20260922T085510
CREATED:20260313T151609Z
LAST-MODIFIED:20260313T151812Z
UID:150608-1774436400-1774441800@www.ri.cmu.edu
SUMMARY:Robust\, Reliable Robot Odometry and its Certification
DESCRIPTION:Abstract:\n\nRobot odometry is the backbone of nearly all modern autonomous systems including\, but not limited to\, unmanned aerial vehicles\, autonomous underwater vehicles\, and autonomous ground vehicles. Most downstream tasks such as path planning\, perception\, and control require accurate knowledge of the vehicle position and orientation at any given moment. While odometry is well-studied and has many potential solutions\, due to its high-importance to the rest of the autonomy stack\, any increase in robustness and accuracy will only further drive the reliability and stability of the given autonomous system.\n\nThere are a number of areas where current odometry methods can be improved in accuracy\, robustness\, or reliability. More specifically\, we find that poor sensor models or assumptions can often reduce reliability and accuracy. Another area of concern is sensor failure\, where odometry methods that are too tightly coupled often fail entirely with a single sensor outage. Finally\, another potential issue is when odometry failures do occur\, most downstream tasks are unaware\, which can lead to erroneous behavior.\n\n\nIn this work we present methods that overcome these challenges. Specifically\, they (1) seek to correct any modeling errors that may occur\, (2) are robust to sensor failure\, and (3) provide sub-optimality metrics for downstream tasks. We first present a method for fusion of wheel encoder measurements for off-road autonomous vehicles that provides piecewise-planar constraints for non-planar environments\, while estimating wheel slip\, wheel radii\, and wheel baseline all in real-time. Additionally\, we present a novel LiDAR odometry method\, with a frontend informed by our empirical evaluations\, and a backend that smooths over a window of prior states while providing map corrections in real-time. This results in more accurate estimates and heightened robustness.\n\nFinally\, we propose finding a fast “sufficient condition” certificate for these optimization-based odometry methods utilizing novel semidefinite programming techniques. While not a perfect catch-all for odometry failures\, it aims to detect when sub-optimality or degeneracies in state estimation may be occurring and pass this information to downstream tasks\, allowing for reactionary behavior.\n\n\nThesis Committee Members:\nMichael Kaess\, chair\nSebastian Scherer\nGeorge Kantor\nTim Barfoot\, University of Toronto\n \nCurrent Thesis Proposal Draft
URL:https://www.ri.cmu.edu/event/robust-reliable-robot-odometry-and-its-certification/
LOCATION:3305 Newell-Simon Hall
CATEGORIES:PhD Thesis Proposal,Student Talks
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260327T110000
DTEND;TZID=America/New_York:20260327T123000
DTSTAMP:20260922T085510
CREATED:20260318T195126Z
LAST-MODIFIED:20260318T195126Z
UID:150634-1774609200-1774614600@www.ri.cmu.edu
SUMMARY:Human-System Communications for Expectation Mismatch
DESCRIPTION:Abstract: \nRobots\, and autonomous systems in general\, are becoming increasingly advancing beyond traditional functions. This can potentially facilitate the mismatch between human expectations of system behaviors during interaction\, especially when the systems behave unexpectedly. Unexpected system behaviors could induce negative emotional responses in humans\, which not all systems have the capability of recognizing and detecting in real-time. To prevent such situations\, systems should communicate system behavior expectations to humans during the task. In addition\, after a mismatch\, the systems should perform post-hoc strategies to mitigate human’s negative emotional responses.\n\n\nThis thesis first investigates how systems can communicate expectations to humans in-situ using legible motion planning based on Potential Field and Vector Field. Such obstacle-aware intent-expressive motion planner could produce comparably legible paths to conventional methods. \nNext\, this thesis explores how systems can detect subtle emotional responses to unexpected system behaviors\, by designing and collecting data from a human study of participants interacting with a driving simulation system to perform non-critical tasks. Findings shown that participants’ emotional responses to different stimuli such as surprise\, confusion\, and frustration\, could be distinguished based on facial action units\, providing important insights to building a real-time autonomous emotional response detector. \nFinally\, this thesis studies a few post-hoc expectation mismatch mitigation strategies. Specifically\, this thesis focuses on acknowledgment\, apology\, and explanation to mitigate negative emotional responses by humans using verbal\, visual\, and nonverbal modalities. Analyses shown that acknowledgment and apology as sole mitigation strategies are not sufficient to influence people’s emotional responses in driving simulator settings\, and further actions from the system are preferred. Moreover\, different modalities of explanation\, such as verbal and visual\, are comparably effective in reducing unexpectedness in robot behaviors\, and such modes of explanations could retain their effectiveness cross-platform. \n\n \nThesis Committee Members:\nAaron Steinfeld (Chair)\nFernando De La Torre Frade\nNikolas Martelaro\nBrian Mok (BMW Group)\n \nDraft of the thesis document
URL:https://www.ri.cmu.edu/event/human-system-communications-for-expectation-mismatch-2/
LOCATION:3305 Newell-Simon Hall
CATEGORIES:PhD Thesis Defense,Student Talks
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20260330T100000
DTEND;TZID=America/New_York:20260330T120000
DTSTAMP:20260922T085510
CREATED:20260320T152243Z
LAST-MODIFIED:20260320T152243Z
UID:150643-1774864800-1774872000@www.ri.cmu.edu
SUMMARY:RI PhD Thesis Defense - Brian Yang
DESCRIPTION:Date: 30th March 2026\nTime: 10:00 a.m. (ET)\nLocation: NSH 3305\nZoom: Link\nType: Ph.D. Thesis Defense\nWho: Brian Yang\nTitle: A Language-Guided Generative Data Engine for Autonomous Driving\n \nAbstract:\n\nAutonomous driving systems struggle with the long tail of rare and safety-critical scenarios that are sparsely represented in real-world data. While recent learning-based planners perform well on common situations\, their reliability degrades in edge cases\, and existing approaches based on simulation or world modeling remain costly and difficult to scale. \nThis thesis proposes a complementary paradigm for scaling driving policies by leveraging language supervision and generative models as a unified data engine. Rather than relying solely on logged demonstrations\, we use vision-language models to infer structured constraints and reward functions from high-level specifications of desired behavior. These programmatic rewards are then used to guide trajectory diffusion models\, enabling the generation of diverse\, physically consistent\, and behaviorally targeted driving trajectories. \nWe first show that combining pretrained diffusion policies with reward guidance enables zero-shot instruction following\, without requiring language-annotated driving data. We then demonstrate how guided diffusion can be used as a scalable offline data generator\, producing counterfactual trajectories that expand coverage of rare and safety-critical scenarios. Policies distilled from this synthetic data outperform those trained purely on logged datasets and no longer require test-time guidance. Finally\, we close the loop by learning reward programs automatically via vision-language feedback\, enabling iterative improvement of both the data generator and the policy. \nTogether\, this work presents a scalable framework in which language specifies what should happen\, and generative models produce how it happens\, yielding a practical path toward robust driving policies without exhaustive real-world data collection or high-fidelity simulation.\n\n\n\n \nThesis Committee Members:\nKaterina Fragkiadaki (co-chair)\nJeff Schneider (co-chair)\nMaxim Likhachev\nPhilipp Krähenbühl (UT Austin)\n \nDraft of the thesis document
URL:https://www.ri.cmu.edu/event/ri-phd-thesis-defense-brian-yang/
LOCATION:Newell-Simon Hall 3305
CATEGORIES:PhD Thesis Defense,Student Talks
END:VEVENT
END:VCALENDAR