Model-based Bio-inspired Approaches for Ego-motion Estimation from Optical Flow

基于模型的仿生方法从光流估计自我运动

• 批准号: 320838831
• 负责人: Professor Dr.-Ing. Kolja Ernst Kühnlenz
• 金额: --
• 依托单位: Fakultät Elektrotechnik/ Informatik
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2016
• 资助国家: 德国
• 起止时间: 2015-12-31 至 2017-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/320838831?language=en
• 关键词: Model; based; Bio; inspired; Approaches

Within the scope of the DFG-Projektakademie "Ingenieurwissenschaften", this project investigates bio-inspired approaches for robotic ego-motion estimation from optical flow. The novelty consists in an extension of efficient, but with respect to estimation quality sub-optimal, insect-eye vision with technical optimization approaches, in order to achieve computational efficiency, accuracy and robustness against disturbances, particularly, stochastical and time delay. Optimization approaches for integration with neuronal models are investigated and robustness of the complete system is evaluated analytically and experimentally. An assessment of the performance with autonomous mobile robots under real-world indoor and outdoor conditions is planned. WIthin the scope of the Projektakademie, this approach is firstly investigated for one-dimensional movements and in a later step generalized to six degrees of freedom within a successor project.

在dfg - project jektakademie“Ingenieurwissenschaften”的范围内，该项目研究了基于光流的机器人自我运动估计的生物启发方法。其新颖性在于扩展了效率，但相对于估计质量次优，昆虫眼视觉与技术优化方法，以实现计算效率，准确性和抗干扰的鲁棒性，特别是随机和时滞。研究了与神经元模型集成的优化方法，并对整个系统的鲁棒性进行了分析和实验评价。计划对自主移动机器人在真实室内和室外条件下的性能进行评估。在Projektakademie的范围内，该方法首先研究一维运动，然后在后续项目中推广到六自由度。