Public research sourceUniversity of PennsylvaniaKumar Robotics / GRASP
UPenn HALO: Finding Drone Targets from Language
First understand what University of Pennsylvania researchers are trying to learn through Kumar Robotics / GRASP, why the question matters, how they test it, and what the result cannot prove. Tell a simulated drone what to find, then compare its route and success when the wording of the instruction changes. The student then completes A drone-search app that resolves ambiguous language with a follow-up question and searches only for an approved target. For a college application or interview, the student separates the source research from their feature, interface, tests, failures, and revisions. Available online or in person in Apgujeong, in one-to-one or small-group formats.
University and lab names show where each public project came from. CIT independently designed these courses; they are not official university courses, partnerships, or endorsements.
8guided sessions
12-24 hoursto the first small project
Advancedrecommended level
Not required for simulationphysical robot
Primary routes
Choose the student's primary AI education and portfolio route
CIT keeps the international-school route first, followed by the separate science and gifted-school route.
CIT lessons can run online across Korea and overseas or in person in Apgujeong. The student owns the question, code, tests and explanation; classes do not guarantee admission, selection or awards, and online delivery never changes an institution's submission rules.
If we tell a drone to 'find a chair,' which route will it take to search for the target?
Study a simulated drone using one camera to find a target named in words. Compare its route and success when the instruction changes.
Students first use a small working example to understand the researchers' question, method, and evidence. A controlled change helps identify what the student's extension must solve. The student then implements a useful feature, connects an operator interface, and tests the integrated application in normal, boundary, and failure cases.
What will the student complete?
A drone-search app that resolves ambiguous language with a follow-up question and searches only for an approved target
The final package includes runnable instructions, the feature and interface design, normal and failure tests, one documented revision, and a three-minute explanation in the student's own words.
Feature
A feature that shows possible targets and asks a follow-up question instead of moving when a spoken goal is ambiguous
Operator interface
A search screen mapping target candidates, drone route, explored area, and the current reason for stopping
Optional LLM boundary
An LLM maps a free-form instruction to allowed target and place candidates and asks for clarification when needed. Only a human-approved target reaches the navigator.
Integrated result
A drone-search app that resolves ambiguous language with a follow-up question and searches only for an approved target
Four ideas explained in this course
01mapping surroundings with a camera
02marking object types on a map
03connecting words to objects
04routes used to find a target
Eight introductory sessions
Understand the research, then build and test a working robotics application
The 20 courses are eight-session CIT studios in which students understand a public university or lab project, confirm a working example, and then turn it into a small robotics application with a useful feature and an operator interface. Selected courses add an optional LLM explanation tool that can read run records but cannot control the robot. RoboMaster has 28 sessions, and implementation scope is adjusted to each student's experience and computer access.
Understand the source
01
Review the simulated drone and safety limits
02
Map the surroundings with one camera
Measure and compare
03
Mark target types on the map
04
Connect an instruction with an object
Build a feature
05
Change only the wording of the instruction
06
Implement the student-owned feature: A feature that shows possible targets and asks a follow-up question instead of moving when a spoken goal is ambiguous
Integrate and demonstrate
07
Build the operator interface: A search screen mapping target candidates, drone route, explored area, and the current reason for stopping. Add the bounded assistant: An LLM maps a free-form instruction to allowed target and place candidates and asks for clarification when needed. Only a human-approved target reaches the navigator.
08
Integrate, test, and demonstrate: A drone-search app that resolves ambiguous language with a follow-up question and searches only for an approved target
Public project used in class
See the exact version CIT reviewed
Public projects can change over time. To keep the class example consistent, CIT uses version eeb361ef02bc of KumarRobotics/HALO. CIT checked it on 2026-08-14; it was created on 2026-07-17. Usage-rights note: No license file found at review; inspect upstream terms before reuse.
licence No license file found at review; inspect upstream terms before reuse
text files 151
The student changes one map setting on an indoor search drone and sees what speed costs in coverage. In class the project is separated into five boxes (input, memory, process, output, control). The lesson opens one of them, changes one value, and leaves the rest closed.
InputWhat comes in?
The forward camera view, and the object a person named in words
Memoryopened hereWhat persists?
The map that cuts seen space into cells and records each as free or blocked
ProcessWhat transforms?
Choosing the next place to go from the frontiers not yet seen, and joining a path to it
OutputWhat leaves, and who uses it?
The path it flew, and the log of where it looked and when
ControlWhat decides when anything runs?
What decides whether it is now exploring, approaching or stopped, and which nodes are running
The files the lesson opens, by name. The course is not a walk through the repository; it opens a chosen few and says which.
One value moves and everything else stays. The right-hand column is the prediction written before the run, not a result; the work is reconciling the two.
The universities and labs named here made the open projects this course reads. CIT designed the course independently; it is not an official, affiliated, or endorsed course.
Clear answers about what students do, what they need, and where the project came from.
Is UPenn HALO: Finding Drone Targets from Language an official course from University of Pennsylvania Kumar Robotics, GRASP?
University and lab names show where each public project came from. CIT independently designed these courses; they are not official university courses, partnerships, or endorsements.
Is hardware required?
Not required for simulation. The course starts with a robot on the computer or a saved recording of a completed run. Computer guidance: ROS environment and GPU recommended.
What background should a student have?
Recommended level: Advanced. Students should be able to follow a guided Python example, test one controlled change, and then build and explain a small feature and interface.
What will the student make?
The completed project is A drone-search app that resolves ambiguous language with a follow-up question and searches only for an approved target. The student implements A feature that shows possible targets and asks a follow-up question instead of moving when a spoken goal is ambiguous and A search screen mapping target candidates, drone route, explored area, and the current reason for stopping. An LLM maps a free-form instruction to allowed target and place candidates and asks for clarification when needed. Only a human-approved target reaches the navigator.
How can the student use this project in a college application?
Explain the source research question, method, evidence, and limits first. Then separate the student's own feature and interface decisions, normal and failure tests, revisions, and next question. The source institution's name does not imply affiliation or guarantee admission.
Which project version does the course use?
CIT reviewed version eeb361ef02bc of KumarRobotics/HALO on 2026-08-14. That version was created on 2026-07-17. We keep this version during class so the example does not change unexpectedly, and we check the setup again before teaching.
Can UPenn HALO: Finding Drone Targets from Language be taken online or in person, one-to-one or in a small group?
Yes. This course is offered online and in person at CIT in Apgujeong, Gangnam-gu, Seoul, with one-to-one and small-group options. Placement and current availability are confirmed after a readiness consultation.
Is this a good first project for this student?
Before placement, we check the student's coding and math experience, available computer, interests, and ability to explain what happened.
Stanford CARLO: Safe Intersection ExperimentsMove cars and pedestrians through a simple on-screen intersection. Change speed and the rule that chooses...→
Harvard AirLearning: Teaching Drones to Avoid ObstaclesSafely study an older public drone-learning project in a current environment. Use saved runs to compare how...→
Stanford robosuite: How Robot Arms Pick Up ObjectsRun a simulated robot arm that lifts an object or opens a door. Change the camera position and movement method...→