Search

Flying insects employ elegant optical-flow-based strategies to solve complex tasks such as landing or obstacle avoidance. Roboticists have mimicked these strategies on flying robots with only limited success, because optical flow (1) cannot disentangle distance from velocity and (2) is less informative in the highly important flight direction. Here, we propose a solution to these fundamental shortcomings by having robots learn to estimate distances to objects by their visual appearance. The learning process obtains supervised targets from a stability-based distance estimation approach. We have successfully implemented the process on a small flying robot. For the task of landing, it results in faster, smooth landings. For the task of obstacle avoidance, it results in higher success rates at higher flight speeds. Our results yield improved robotic visual navigation capabilities and lead to a novel hypothesis on insect intelligence: behaviours that were described as optical-flow-based and hardwired actually benefit from learning processes. Autonomous flight is challenging for small flying robots, given the limited space for sensors and on-board processing capabilities, but a promising approach is to mimic optical-flow-based strategies of flying insects. A new development improves this technique, enabling smoother landings and better obstacle avoidance, by giving robots the ability to learn to estimate distances to objects by their visual appearance.

After finishing a traditional (theory/tool-orientated) class on PM, students do not necessarily understand the connection between the phases and the overall need for PM. Therefore, a game was designed to let the students experience and reflect the challenges of project work. The game is an adaptation of the “egg-dropping-challenge” (Cp. e.g. http://eggdropproject.org/). The challenge is to create a structure for a raw egg that protects it from cracking at a specific height of a fall. This challenge is in my version integrated in a comprehensive project cycle. The pen portrait describes the idea of the game and the implementation.

Desert ants Cataglyphis fortis have been shown to be able to employ two mechanisms of dis- tance estimation: exploiting both optic Xow and propri- oceptive information. This study aims at understanding possible interactions between the two possibly redun- dant mechanisms of distance estimation. We ask whether in Cataglyphis the obviously minor contribu- tion of optic Xow would increase or even take over completely if the ants were deprived of reliable propri- oceptive information. In various experimental para- digms ants were subjected to passive horizontal displacements during which they perceived optic Xow, but were prohibited from active locomotion. The results show that in desert ants active locomotion is essential for providing the ants' odometer and hence its path integrator with the necessary information.

Here we report on a validation study on brain–machine interfaces (BMIs) performed during the December 2007 ESA parabolic flight campaign. We investigated the feasibility of using BMIs for space applications by performing tests in microgravity. Brain signals were recorded with noninvasive electroencephalography before (calibration sessions) and during the parabolic flights on two subjects with prior BMI experience. The results of our experiments show that an experienced BMI user can achieve stable performance in all gravity conditions examined and, hence, demonstrate the feasibility of operating noninvasive BMIs in space.

Insects exhibit remarkable navigation capabilities that current control architectures are still far from successfully mimic and reproduce. In this chapter, we present the results of a study on conceptualizing insect/machine hybrid controllers for improving autonomy of exploratory vehicles. First, the different principally possible levels of interfacing between insect and machine are examined followed by a review of current approaches towards hybridity and enabling technologies. Based on the insights of this activity, we propose a double hybrid control architecture which hinges around the concept of "insect-in-a-cockpit." It integrates both biological/artificial (insect/robot) modules and deliberative/reactive behavior. The basic assumption is that "low-level" tasks are managed by the robot, while the "insect intelligence" is exploited whenever high-level problem solving and decision making is required. Both neural and natural interfacing have been considered to achieve robustness and redundancy of exchanged information.