LPLeo Pape

Leo Pape

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Leo Pape has a total of 9 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Learning tactile skills through curious exploration

    by Leo Pape, Calogero M. Oddo, Marco Controzzi, Christian Cipriani, Alexander Förster, Maria C. Carrozza, Jürgen Schmidhuber

    We present curiosity-driven, autonomous acquisition of tactile exploratory skills on a biomimetic robot finger equipped with an array of microelectromechanical touch sensors. Instead of building tailored algorithms for solving a specific tactile task, we employ a more general curiosity-driven reinforcement learning approach that autonomously learns a set of motor skills in absence of an explicit teacher signal. In this approach, the acquisition of skills is driven by the information content of the sensory input signals relative to a learner that aims at representing sensory inputs using fewer and fewer computational resources. We show that, from initially random exploration of its environment, the robotic system autonomously develops a small set of basic motor skills that lead to different kinds of tactile input. Next, the system learns how to exploit the learned motor skills to solve supervised texture classification tasks. Our approach demonstrates the feasibility of autonomous acquisition of tactile skills on physical robotic platforms through curiosity-driven reinforcement learning, overcomes typical difficulties of engineered solutions for active tactile exploration and underactuated control, and provides a basis for studying developmental learning through intrinsic motivation in robots.


  • Models and scales for cross‐shore sandbar migration

    by L. Pape, Y. Kuriyama, B. G. Ruessink

    We investigate the long‐term (months to years) predictability of cross‐shore sandbar migration with two models that operate on different abstraction levels: (1) a coupled, cross‐shore waves‐currents‐bathymetric evolution model and (2) two data‐driven neural network models, based on simplified cross‐shore profile representations and daily averaged wave properties. For model calibration, training, and validation, we use a high‐resolution 15 yearlong profile data set collected at the Hasaki Oceanographic Research Station in Japan. Sandbar behavior at this field site is characterized by cycles of net‐offshore migration with a duration of 1–4 years. We find that all models can produce several general features of sandbar behavior at the studied field site, such as rapid offshore migration, slower onshore return, and net‐offshore migration. However, it is difficult to quantitatively predict the offshore‐directed trends in sandbar location over time scales of months to years. While simple linear models outperform more detailed nonlinear models, for all models it is difficult to predict long‐term sandbar behavior, because of error accumulation in the model's processes over time. Representing processes on a more abstract level (scale aggregation) alleviates error accumulation but does not completely overcome this problem.


  • Open Access

    On cross‐shore migration and equilibrium states of nearshore sandbars

    by L. Pape, N. G. Plant, B. G. Ruessink

    The location of submerged sandbars in the nearshore zone changes over time in response to variability in wave conditions. Here, cross‐shore sandbar migration is studied with an empirical model consisting of a differential equation relating cross‐shore sandbar migration to wave forcing. Model parameters are fitted to data sets containing multiple years of daily‐observed positions of five sandbars at three field sites. From the fitted model parameters we determine the cross‐shore location of zero‐migration (equilibrium location), the stability of this equilibrium, and the rate (response time) at which a sandbar migrates toward or away from the equilibrium location. We find that for breaking waves a sandbar moves toward a stable, wave‐height‐dependent equilibrium location. For non‐ to slightly‐breaking waves, however, a sandbar moves away from the equilibrium location, implying that the beach profile evolves toward a state without submerged sandbars. The response times can differ up to a factor three for different sandbars at a particular field site, and a factor ten between sandbars at different sites. For breaking waves, response times are found to decrease with increasing wave height from months to days, while for non‐ to slightly‐breaking waves response times exceed several months. In general, response times remain larger than the timescale of variability in wave climate (several days), implying that sandbars spend most of their lifetime out of equilibrium with the wave forcing. The model is able to hindcast the relevant features of cross‐shore sandbar migration on timescales of days to several years, and outperforms a baseline prediction of no change for all sandbars. When the equilibrium location is reached by both the hindcasted and actual sandbar, the hindcast error depends on the accuracy of the predicted equilibrium location only, and no longer on the history of accumulated errors. As a result, sandbars that occasionally reach their equilibrium location can accurately be predicted from the wave forcings over their entire lifespan. However, it is difficult to predict the exact rate of the yearly to interannual trends that are observed at two field sites, because the sandbars at these sites never reach the offshore‐located equilibrium states associated with high waves.