I can offer UROPs in the areas listed below. Relevant keywords listed after each project describe the skills you will need for each project. Some projects may lead to paid UROPS over the summer.

• Robot software development

  • Developing software for autonomous helicopter path planning (Linux, C, Graphics)
  • Developing CARMEN interface libraries (C, Java)
  • Developing robotic test protocols (Linux, C, Robotics)

• Human-computer interaction

  • Dialogue Management for Flo, the Nurse-Robot (Linux, C, Perl, Graphics)
  • Measuring Age Dependent Response To Collision Avoidance Systems In The Automobile. (Visual Basic, some hardware)
  
  

  ---

  
  

• Developing software for autonomous helicopter path planning.

  There is a helicopter simulator called WinHeliSim that currently exists only for Windows. The initial part of this UROP will be porting WinHeliSim from Windows to Linux. The major part of this effort will involve GUI programming. The second part of this UROP will be implementing path-planning algorithms on the helicopter. If you are successful, we will test the algorithms on an actual autonomous helicopter.

  You will need thorough knowledge of C and Linux for this project, and a comfort level with hardware. Additionally, you will need experience in GUI programming, preferably with GTK+. However, I would not expect you to be an expert in helicopter control but would help you learn the necessary fundamentals.

• Dialogue Management for Flo.

  Flo is the original prototype of the Nursebot, a personal robotic assistant. The Nursebot project is an inter-disciplinary multi-university research initiative focused on robotic technology for the elderly.

  This UROP would initially consist of updating the existing dialogue manager and integrating the dialogue manager with robot motion control software. Once the existing dialogue manager was operational again, the second phase of the project would consist of implementing and assessing new generations of dialogue management algorithms.

  You will need thorough knowledge of C, perl and Linux for this project, and a comfort level with hardware. However, I would not expect you to be an expert in dialogue management but would help you learn the necessary fundamentals.

• Measuring Age Dependent Response To Collision Avoidance Systems In The Automobile.

  (In conjunction with Bryan Reimer.)

  Collision avoidance is fast becoming one of the new "hot" technologies integrated into the automobile. BMW and Daimler Chrysler have incorporated collision avoidance systems for parallel parking into some of their new vehicle lines. It is likely that in the very near future similar systems will be incorporated into vehicles to detect long-range high-speed collisions. Little is know about the effect of these systems on different age groups. There is limited data showing that older drivers are less likely then younger ones to trust this type of technology. Furthermore, there is growing concern that the effect of these systems will be much like the effect of anti-lock brakes introduced as a safety device in the 80's. Older drivers who do not understand the response of the anti-lock breaking system often improperly apply the brakes in an emergency, where as younger drivers are more confident of their increased deceleration and thus more likely to driver faster. The project will require some background research on the effect of anti-lock brakes on the entire driving population. Experiments on collision avoidance will be conducted in "Miss Daisy" the MIT Age Lab driving simulator to quantify the effects of cognitive load, false positives and presentation medium. A testing approach will need to be developed to address these key variables over different age cohorts.

  You will need to know (or be willing to learn) Visual Basic for this project, and be comfortable working with simple hardware. Dr. Reimer and I will not be comfortable working with a single student on this project, but would prefer a pair of students. If you are looking for a project to do with a friend, this may be a good one. Furthermore, there is a possibility of this project extending into the summer for pay.

• Developing CARMEN interface libraries

  CARMEN is a robot control software package developed between CMU, Stanford and MIT. The software is fully distributed, and is capable of running across multiple machines. The CARMEN platform provides programmatic control through interface libraries, but these libraries are available only in C. I would like someone to write a Java API and be able to control the robot with a pure Java program.

  You will need a reasonable knowledge of C and Java for this project. You will need to be familiar with Linux, but not an expert. You will not need to know much about robotics, but you will hopefully learn as you go.

• Developing robotic test protocols

  As described by the previous project, CARMEN is a robot control software package developed between CMU, Stanford and MIT. The software is fully distributed, and is capable of running across multiple machines. The CARMEN platform provides programmatic control through interface libraries, but these libraries are available only in C. I would like someone to write a Java API and be able to control the robot with a pure Java program. Unfortunately, there is no test harness for CARMEN. When the software is updated, there is no way to verify the performance. I would like a student to write a test harness and a suite of tests for automatic testing of the software.

  You will need a reasonable knowledge of C. You will need to be familiar with Linux, but not an expert. You will however need to know a reasonable amount about mobile robotics.